 Thank you for joining us everyone. We're just going to give it a few seconds to let people log on before we kick off formally. I can just see the numbers going up so we'll just wait a few minutes for everyone to get comfortable and ready for today's event. Ond bod yr enw yn gweithio tra'w amser gyda'r tynnu allan, rwy'n cael ei dweud i'r papyr rhwng diwrnod Cymru ar y D�rymenau Paul-Mellon yn y Gymhelfarad Feel. Rwy'n gweithio i'r Llywodraeth Microedd Maxxer, ac rwy'n gweithio'r Llywodraeth Pooh Llywodraeth Loed. Rwy'r llunau Ll來 yn ymgyrch yn roi ddiolch. этиb yn Gymryddon Birfyddon, ddodd Llan, gyda ddiolch ar y llunion Ond rwy'n gweithio'r llunion gan eich cyfnod yma yn gweithio i ddechrau o'i gygu â'r ffordd ar y cyfnod'r website. Maen nhw'n gweithio i ddim hanfod i byw o ddweud â'r website,fa ychydig o'r ddweud yn oed fel yna yn mydd y centre. Sfer y pwynd i yw'r gweithio, mae gwirio 你wch yn gwneud drwy gyflwyno a'r lookson, rhaid i hefyd y Llyfridd. Rwy'n meddwl am ddif� am y ddechrau, mae yw gweithio ar gyfer y cyfrwno, rwy'n hynod dweud y rhaid i ddweud, ac i ddweud y gwrthodau i gael ei gyrdd, ac mae'n gweithio'n gweithio'n gweithio hynny, rwy'n ddweud o'r digwydd ac mae'n gwneud eich cymryd, i ddweud gan amlwg i'r gwneud ac i'r gweithio i gael gwneud, a'r gweithio i'r eventau yma yn ysgrifetol dechrau, i'r reisio'n gyfathodol, i'r ddau'r gweithio'n gweithio. Lleiddoedd hi wnaeth y werth ynglyn â'r pethau yn byddwyr, ond yno, Martin, Michael Oeddon, Mae Co-connut yn gynnyddio'r gymuned yn gynllun o'r genesys gynnig o'r gyda chi ammellai ac addysg ar gyfer bod ychydig iawn a gwnaeth arnynach ffordd o ddau'r methodologau. Ond o dddarparu hwn yn eich bod ni'n gychwyn o gydig iawn i ddeudio ddyliadol, o ddau'r ddau ffysg – o ddiddoriaeth, arddurwch, storiograffi, a methodologi. Rydyn ni'n dweud bod yw'r ffordd yn y ffordd o'r ysgrifennu ar y technologi ac methodologi, a'r ysgrifennu o'r pethau o'r ddweud o'r ddweud o'r ymgyrch yn y rhan o'r gweithio, ymgyrch, ymgyrch, ymgyrch, ymgyrch, ymgyrch. Yn ymgyrch ymgyrch, mae'n cyfrifio, fel ymgyrch o'r ddweud o'r prynch, a'r ddigital, ymgyrch o'r ddweud o'r ymgyrch o'r ddweud o'r ddweud o'r ddweud o'r ymgyrch o'r bethau o'r ddweud o'r prydian ystafell o'r ddweud o'r ddweud o'r ddweud o'r treidog o'r ddweud, ad roeddwn i'n gwagoroddau cyfry, ar ddweud y ddweud o'r ddweud, ac yn dweud o'r ddweud o'r ddweud o'r ddweud o'r sgwpwy功wll ymgyrch o'r ysgrifennu ar yr ardydd. Ac derbyn conversation iddyn nhw, journal and open access, born digital journal British Arts Studies and you can find some articles there which are exploring mass data issues. And also our Royal Academy Chronicle and Danny is going to put the links to those publications in the chat and I know that one of our speakers is also going to use the data from the Royal Academy project as well so it's great to see the afterlife of that publication and the thinking around that. It's been spoken about today. So we really hope that like I say today will generate as many questions as it answers and think about further avenues and connections for thinking through these ideas. And before we get on to more of the the sort of meat of the intellectual inquiry of today. I just want to quickly walk you through some of the housekeeping that we have for Paul Mellon Centre virtual gatherings to make sure that these are really enjoyable and productive spaces for everyone involved. And you can see that as audience members, you can really interact with the event, and we really encourage you to be active in the chat box I can see that you're already interacting with us which is great so hi to everyone in the chat box as well. But if you want to put formal questions at the after the papers, please type those in the Q&A box and the chairs of our panel will read those out to the speakers. And we also do have a function by which you can speak to us and we can hear you. And you can put your question live, as it were, by speaking, if you'd like to do that, you use the raise hand function, and we can then unmute you and ask you to join the conversation. The session is being recorded and made available to the public. And again, I'd encourage you to go and have a look at the recording section of the PMC website. We do try and record a lot of our events and they're becoming their own kind of online resource and publication in their own rights. So it's there are great resource to point to your students if you teach or any of your other networks as ways of again interacting with events that you might not have been able to participate in or watch live. And you can also use a CC function to enable captioning, and you can find that at the bottom of your zoom screen. And again, like I say, to make sure this is a safe and productive environment. We will we do reserve the right to remove any audience members for offensive behaviour, but we really hope that doesn't happen and that this is like I say a really productive and generous space for everyone involved. So without further ado, I'm going to hand over to my colleague and co convener, Martin Myrone to say more about the topic and how we got here today. Hi, Martin, over to you. Hello everybody. Yes, I'm Martin Myrone. I'm head of grants fellowships and networks here at the Paul Mellon Centre. And with Sarah Turner and Shreya Chatterty. We've kind of worked on on framing today's event and sending out the invitation. And well, before anything else, so thank you to everybody who's contributing today, and everybody who's listening in or watching in as well. And even if you're not going to use the chat to post comments that's worth looking in on because I've just been seeing comments appearing from all around the world and Rome and Cairo. And this, yeah, it is, it is an international participation, which is always fantastically heartening to see and a reflection of the digital world that we're in, which, you know, for what its limitations is also hugely enabling as well. As Sarah's already kind of indicated, we've been thinking about this event or something in the area which would lead to this event for quite a while now. We've been kind of checking back through email exchanges and discussions. It was really the winter of 2020 to 21, where, in a kind of those kind of happy coincidences we had several different conversations, which seemed to indicate that an awareness that out there, and a whole range of different contexts with people working in potentially very different ways, and in different environments into different ends, when nonetheless engaged in research or book projects, which we're dealing with and representing large volumes of data in order to meet some historical objective, or to address an area of inquiry. And there was a sense that there was, is there something new, is there something novel, is there something interesting, is there something that's worth reflecting on in that range of work, accepting this diversity, that there may be some common ground or some common questions. And I guess there were probably a couple of points which occurred to us when we started with that observation. It was that a lot of this work was focused on objects and on the records of objects and the way objects have been circulated and recorded, notably in the art market. But there were also projects that were looking slightly differently at populations, at people, kind of human data, bodies of data relating to kind of human entities, artists and consumers, and that seemed to be kind of interesting. So that was one thing to note. The other thing was a sense that this wasn't digital, it was digital art history, but it wasn't something that was contained by the idea of digital art history. The outputs might be digital, and we've mentioned a couple of projects already which have been published online. There were also book projects, there were people who were working towards book, and people who were working towards kind of quite established art historical perspectives as well. So technology was enabling, technology was facilitating, technology was the medium, but it wasn't an end of itself. There was something there which wasn't being addressed in the discourse around digital art history, but was more about the intersection between technology and mass data, and art historical practice in a broad sense. And in that regard, I guess it's notable that there are presidents for the sorts of work that we're recognizing, and thinking needs to be represented and connected up, which were distinctly non-digital, or which were developed before modern mass data technology, and Montias on Dutch art, or White on White on 19th century France. I'm sure they may have had some access to rudimentary databases, but they were producing books, and they were working for some material which perhaps are massed and analysed in ways which preceded the sorts of technological instruments that we have now. And we are aware, kind of broadly aware that there was a whole range of research questions that arise from this about how we think about artists as a group, and how we think about previously neglected source materials which were newly available including exhibition catalogs, commercial directories, census records and news and media, and all of which can be interrogated much more readily and in potentially very revealing ways. And out of that comes a question about questions around what does it mean to think about industry beyond the singular, to think about the mass rather than the individual about patterns and populations, and what does it mean to think not about individual objects, but objects in large volumes. And what does it mean to think about artistic populations and how these intersect with larger political entities like community class and nation. And that led to an exploratory session in May 2021, which included a number of people here here today, but also Paris Spies Gans and whose book on women artists being, which was due to be published by the Paul Mellon Centre, which is now out, was one of the projects of which we thought was interesting and that it was, you know, it was a book project but it was utilising large volumes of information. And that workshop really kind of just kind of looking at what what work was out there and what the potential was led eventually to the call for papers and to today's session. The questions which are kind of very much on our mind today as we hear the presentations are was a fundamental question rather brutal question. Why do this work. What's the point. What impacts does what impact does mass data have upon our historical methodologies. What does it mean to have this data, and what impact does it have on art history as a discipline. But most of all, and this I think is the kind of endocrine which emerged out of that exploratory workshop and which we're thinking about today. What can we learn from each other. What can we learn across disciplines, across methodologies, across national and cultural contexts, across different kinds of material and different periods. And although we know we as the Paul Mellon Centre are very much focused on British art and that's our raison d'etre, and that's what we're we're aim for we have a very open and porous idea of what British art is. It's a place to start asking questions rather than in itself. And today's activities have been framed by without reference to British art as a limiting notion. There are people addressing British art, but we're really interested in creating international connections and facilitating a kind of comparative dialogue across disciplines across fields in order to, we hope, enrich and deepen our understanding of mass data methodologies and art history. Hence the range of papers and the range of presentations that we have today which cover. I think quite a kind of dazzling array, potentially dazzling array of materials periods methodologies and kinds of work. But without further ado, we should move into panel one. The format for today is to have quite succinct papers but they've got 15 minute slots for each presentation. A number of presentations are collaborative they have more than one speaker involved, and we will run through the papers in sequence. We will very briefly introduce the speakers were full of biographies online, and then have a more sustained Q&A session to take us up to the lunchtime break where we'll reflect and go into conversation around the around the presentations together. So panel one, we're going to start with Anita Gowers and Paul Wilson. Anita Gowers has worked in the higher education sector across a range of disciplines from neuroscience to the creative creative industries and is currently undertaking an artist PhD on the Australian picture framing industry at the Australian National University. Paul Wilson focuses on applying technologies to advance our individual and collective thinking and performance, and it's produced a whole range of data analysis and visualizations including large historical data sets. We have postgraduate courses on technology, leadership and data science. And together, they are addressing, well this is a pretty fundamental question, why art history needs data science. So, over to you, Anita and Paul. Thank you very much. I will share my screen. Okay, just confirming you can see the start of the presentation now. That's great. Thanks, Paul. Thanks, Martin. Our purpose today is to share some brief insights into what data science is about, how you can use it effectively, and the implications for art history research. To begin with, I'm not an art historian myself. I've worked across fields, including a range of data science projects over about 20 years. My co-presenter, Anita Gowers, will explore some analysis about history data in a bit more detail. I'm going to briefly explore some origin ideas and approaches to data science. The basic story of data science is about, it comes from our limited mental capabilities and the discovery that we can use things outside ourselves and actually need to to advance the way we think. There's a lot of recent research from Nobel Prize winners like Daniel Kahneman and Richard Taylor that was recognised the limitations of our thinking and the importance of being able to leverage the things that are part of what's called the extended mind that's out of the work of David Chalmers and Andy Clarke from the late 1990s. This is a really interesting, exciting area. I'd encourage you to have a bit more of a look into it, but I won't go into much details now. The key thing is that the massive growth of technology has really changed what's possible. Businesses and industry are really getting involved in taking on what's called data driven decision making and the World Bank and others are talking about a fourth industrial revolution. The activities going on here are part of a much larger puzzle and also means there's going to be interesting feedback between different fields of human endeavour as different fields pick up and really start to accelerate with these sets of ideas and technologies. In terms of practical points, I'd suggest that you routinely use data visualisations to review your data and findings. Most of our brains are dedicated to processing visual information and it just makes it much easier to see there's some glitch in your data or even something that was unexpected that leads to a new insight. Tables and numbers just don't work and for that reason I strongly encourage you to hunt anomalies until you resolve them. You never know where they lead. I've discovered glitches in the Australian Bureau of Statistics data and been able to confirm that they've got an issue there. So not making assumptions about the data is important too. I anticipate that data science will be new to many people in the field of art history research. In these circumstances, a particularly nice method to use is topic modelling or latent Dirichlet allocation. This is essentially based on the work of David Bligh. It provides a nice way to discover the thematic structure in large collections of documents and to help visualise them. It's nice because you don't actually need to put a structuring to the text. This process will start just classifying things without you needing to have any initial ideas. I've used this process for about 10 years. It involves a little bit of work to get the documents ready, but it's very versatile and you can build up on it in various ways. So the example of David's work I've got here is classified the content of Science magazine articles in decades. You probably do the same sort of thing with information about different artists across time to get just a bit of a flavour of what's happening. It's not about presenting particular answers but exploring the themes that might be worth delving into in a little bit more detail. He's also used this to analyse the reading habits of US politicians and figure out a bit more about... Paul, I just want to check. I'm not seeing the change of your slides and I think I'm just looking in the chat box and other people. There we go. So that's David Bligh and that's the analysis that I was talking about there. He's got a very interesting presentation on YouTube, probabilistic topic models, and it goes through some of these in a bit more detail. I'll keep moving along. I'm not sure, yes. I work for a company that had $40 million to help Tasmanian business and industry adopt e-commerce at one stage. And a lot of business and infrastructure applying for funding. And what I found was that there's somebody else in the world who had already done what this particular business industry was thinking of. What that means is you don't need to consider that you're doing something new and unique. You can test out ideas by just looking around to see what people have done. It's almost like you're a research lab without having to spend any money or effort. One of the sites that's useful for this is flying data. They've got a subscription service and various explanations or tutorials that we've actually examples of what you can do. It's an alternative. I'd suggest that for people who are getting into the field, you don't necessarily need to leap into the details yourself. You might be able to find a young graduate or postgraduate student and have them pursue some of this for you. Another interesting group is Kaggle. They have a large community of data scientists. They have data sets and they have notebooks which are full explanations or code to do different types of analysis. Another interesting thing they do is have competitions. This is where a company provides a data set in some goal and perhaps a financial reward. What happens is they reward the data scientists to contribute to a solution. This is nice because you don't need to understand the technologies involved or the concepts involved. You can get some very smart people working on the challenge that you're trying to understand without needing to be an expert yourself. Just on a final note, the movie Moneyball is quite an interesting example of what can be done with data analytics. The basic story is that talent scouts, however long the sport had been played, had been focused on developing or identifying people to recruit. A Yale economist and a coach realised that they could apply data analytics methods and get high-performing teams for much less money. I found it quite interesting that the author of this work, the researcher, said that a lot of the human experts that were involved in this scouting activity were focused on the really flashy aspects of the player's behaviour, like being able to hit or run and this sort of thing, but there was really important other things that were more subtle than had been missed entirely. I think that provides scope for thinking that maybe there's some things in the artistic side of things or in understanding great works that have been overlooked for some subtle reason that might not be apparent. In conclusion, I think one of the interesting things is the technology convergence that's going on. There's analysis methods that could put on a platform for one domain of human knowledge can equally be applied to others. In the same way, career paths are converging. Data science professions are in demand, so just as the nature of the work means you're not locked into art history. That's probably a good thing because there's so many ideas that can be drawn from elsewhere. There will be pathways into and out of art history and that I think will fuel a positive cycle of growth. It's also worth noting that art history is heading into interesting competitive territory. We've got a lot of focus on having people's attention and allowing them to pursue their interests in a useful way. This is a bit of a challenging area because something else is a mouse click away. It's really important that offerings are compelling, engaging and allow some sort of interaction I'd suggest to really help people feel engaging and get value. Finally, if you're starting out your data science project for the first time, remember it's fine to start small. There's no need to take on the world. It's actually a lot of fun to create analysis insights that people haven't seen before. They get excited and new or two. Even a small success makes the next step easier to take. Thanks very much. I'll hand over to Anita. Onda, share my screen. Right. Can everyone see my slides? That's great. Thanks, Anita. Terrific. And thank you very much, Paul. And hello everyone. As mentioned, I'm Anita and I'm a PhD candidate in the Centre for Art History at ANU. Let me begin by saying up front, I am not a data scientist and I am not a statistician. So what I'm presenting today is basically the useful things that I've picked up about how to interrogate art history data using scientific methodologies. So in other words, I'm giving the practical to Paul's theoretical provocations. My returner generously shared a subset of data behind the online archives, the Royal Academy summer exhibition, a chronicle 1769 to 2018. It's a longitudinal data set. And whilst there are a couple of patchy areas of the data, it is still considerably a remarkable timeframe for a data set that is found within the discipline of art history. The information on this website is presented in depth and annually. So here is a cut and paste from some of the slides that are online on Chronicle 250 displayed from 1769. And what we can see here depicting the exhibitors of Royal Academy members and non Academy members as a percentage within a pie chart. And on the right hand side, it's the number of works displayed and that includes all works again by the Academy members and non Academy members. So what is really clearly visual with these pie charts is that for 1769, there were far more academicians compared to non academicians and logically there were more works by members of the Academy. But what I'm interested in is examining this subset to see if there are longitudinal trends that span the entire data set, which will potentially generate questions about the data. So I've received this data in an Excel file format consisting of 17 variables and realise relatively quickly that mapping each of these would be incredibly time consuming. So we decided to pop this information into a scatter matrix. Don't adjust your eyesight. I have done the font small deliberately. So essentially a scatter matrix is a set of points plotted on a horizontal and a vertical axis to give a summary of all of the relationships between the variables. It visualises these pairs allowing many relationships to be explored within the one chart within the one single visualisation. And what the diagonal shows the distribution of the data, the data above the diagonal plots the inverse of the data below. So in other words, it swaps the vertical and the horizontal diagonals. And with a quick glance, a few things jump out. We can see the remarkable similarities between the number of academicians and the number of exhibits because they track nicely together. However, there's something interesting going on within the exhibition length, which is up here and as well as the attendance, which is here. And also the sales because you would assume the longer the exhibition, the more people and the more sales. So to look at that data further, I've popped it into a bar chart. So this is the exhibitors with academy members and non academy members. So again, don't adjust your eyes. The point of this bar chart is for your eyes to look for the anomalies in the data without focusing on the numbers. So you can see this bar chart displays the academicians, which is the blue, which is remarkably consistent over time, with the exception of the very first year being 1769. And the non-academicians, we can see some significant variations here 1771. There's a drop in 1805, sort of 1850 to 1860 here. There's some high points, a drop here, which is First World War, Second World War, and also the Brenton Woods system in the early 1970s. If we take that same information and drop it into what's called a percentage stacked bar chart, this time we can see the 1769 data, which was barely visible here, absolutely pops out at us in this demonstration of methodology. 1771 also pops out and you can see some of these other anomalies as well. So the display of the data impacts how we mentally perceive it. I've done the same thing here with exhibits and year showing the non-members in red and the Royal Academy members in blue. And again, you can see this remarkable variation. So as art historians, you'd be looking at these variations to then go back to the data, say, well, why are there these peaks? Why are there these troughs? And potentially, on a very broad scale, we could say, well, maybe there's a much stronger relationship between non-academy members exhibiting depending on what's going on in the broader world. These gaps here represent the gaps in the data. And when we pop this exact same information back into the stack chart, we can also again see how it shows us different things, particularly this one, which wasn't really noticeable in the former chart. Now, this visualisation is a scatter plot and that shows the relationships between two discontinuous variables. We can see clusters and each dot on the scatter plot represents a pair of exhibition length and attendance. So intuitively, you would think that the longer an exhibition, you would achieve greater attendance. But what the data shows is that over the last 250 years, exhibitions that lasted approximately 90 days achieve the maximum visitor numbers. And exhibitions longer than 90 days didn't necessarily increase attendance. Again, whilst this scatter plot plots greater sales, they're correlated to exhibitions. So what we can see is that 70 days is basically the medium length. So the correlation is not causation. So whilst there are relationships between the variables, one doesn't necessarily cause the other. And in both of these scatter plots, they are only comparing two variables and there are clearly other things at play. So in other words, making an exhibition in length 70 days is not necessarily going to achieve more sales. But when data is presented in using these scientific methodologies, it helps you ask more questions and refine what you're looking at. So by asking questions of the data drives new discoveries and by looking at the data using scientific methodologies, it helps contextualise art history for us and it actually makes it more relevant for those outside of the discipline. It is not a panacea for art history research. It is not a conveyor belt process where you enter the data, convert it to visuals and suddenly you have the answer on your PhD. Data needs to be tuned. It's a process of exploring, looking, examining and checking that will generate new questions and new research. So data science is a different way of understanding our contribution to the commonwealth of art history knowledge. And it is imperative that we as art historians adapt these new data science techniques alongside traditional art historical methodologies in order to adopt new generations to explore art history. And I hope this has been helpful in showing how visualising data in our discipline provides new avenues potentially with new future research questions. Thank you. Great. Thank you so much, both of you, for a really kind of visually stimulating as well as kind of intellectually stimulating start of the day. So thank you for that. Yes, it's a little bit difficult quite sure what to do with applause, applause, you know, there are smiles, you can put thumbs up. But also, if you've got questions, observations, comments, we will take them together in the Q&A session. But I think you can post questions at any time in the Q&A box. It's quite, you know, if something is occurring to you, do take the opportunity to put it down now. Otherwise, we'll turn to them in the group discussion after our next two presentations. And we'll move on next to Shane Marcy, who's a PhD candidate in the Department of Art, Art History and Visual Studies at Duke University. And Shane's presentation today is digital humanities as a fundamental methodological approach. A case study of the postcard phrase, which I understand it relates very closely to your doctoral project, although I know you're also a active author on the online dictionary of art historians. So I think things kind of connect together over the morning as well. So over to you, Shane. Great. Thank you very much. Can you hear me? Yep, that's all good. That's great. Thank you very much. And thank you Anita and Paul. That was fascinating, especially what Paul mentioned about the importance of starting small and maybe even the pathway out of art history that you mentioned. But so I won't reiterate what you've just said in your introduction, but I am involved in a number of digital projects and the project manager of digital public buildings, North Carolina, but I do also participate in the dictionary of art historians. I want to do other work at the Colab at Duke around 3D printing and trying to expand the parameters of, of what can be done with that as a medium. Sorry, just sharing my screen there. Right, can you see that now? Excellent. That's good. A particular project is distinct for a number of reasons. For a starter represents, as you mentioned, a component of my own dissertation, but it's also a case study on the basis of which I want to suggest a kind of broad but relatively simple argument in favour of mass data methodologies as a kind of fundamental aspect of research in the humanities. It's a kind of rudimentary digital research as a kind of basic form of best practice based on the accessibility of a wide range of tools and the increasing availability of digitized data, especially in the form of online archives. And the capacity that these resources hold for carrying out kind of manageable projects that are limited in scope, but which have the potential to overturn. So, my dissertation just in brief looks at American modernity through the lens of the postcard craze, which took place between in the United States between about 1907 and 1920, during which period the nation was inundated with these objects. And I explore the medium as both an object of art historical investigation, but also as an exemplary form of mass produced commodity in order to ask questions about, in particular, the political economy of the progressive era, I think maybe the best way to just succinctly describe my kind of evolving approach to the medium is to look at the relationship between both sides of the object. On the one hand, on the one side, and through an exploration of the inscribed messages, I ask how the postcards insistence on personal interaction was able to establish a uniquely intimate engagement with what was often a kind of quite alienating mass produced commodity form. On the other side of the object, I'm looking at the image, especially the view card and exploring how it adapted new technologies photography, which was accorded a kind of scientific verisimilitude, but also chromolithography, which carried profoundly commercial connotations, and how these came together through the scale of the postcard craze to create a kind of totalising and commodified vision of reality. Maybe what Walter Benjamin might describe or Marx might describe as a phantasmagoria. And ultimately, I argue that in putting the two sides together, that the intimacy created by the personal engagement with the commodity form allowed for a far stronger kind of identification with this phantasmagoric reality that's suggested by the medium's imagery. And that in doing so, the medium helped to establish a very particular and important form of identification between the modern individual and the commodity form. And significantly, this took place at a time when this was precisely what the American advertising industry was striving for. And of course, you know, this kind of relationship continues to exert a powerful force in our own everyday experiences, but that's the project overall. The digital component of this was concerned with something far more specific, which was visualising the scale of the craze. Specifically, the numbers of postcards purchased in the United States. So, in terms of this project, this is primarily a production analysis. And it's crucial because part of the argument is that these images were produced in such quantities and were so ubiquitous that they impacted the way in which people perceived their everyday experiences. So in terms of the phenomenon scale, both the scholarship and the primary sources clearly convey a narrative of this explosion in postcard sales alluded to in this image of Uncle Sam drowning in a sea of postcards followed by a complete crash. So what I did first of all was to gather information in order to verify this assumption. And so I said about visualising the existing data that is the data for the postcard craze that is currently cited in the literature. And this is based on a book by George and Dorothy Miller upon which almost all of the subsequent scholarship relies. The data themselves are drawn from the Postmaster's General Annual Report, which is a government document that has been made available online in recent years. And what the millers do is that they cite the data for the years that they believe represent the peak of the craze, which for them is about 1908 to 1913. So this is a very small data set that they're using and it looks like this when it's visualised. And on the basis of this data, the millers determined that the peak of the craze occurs in about 1913 when about one billion postcards were sent annually. So in short, what the millers did was they decided they knew the dates of the craze. They went and got the data to back that up at a time when this data was far less accessible. And since then, nobody has kind of critiqued this methodology in spite of the fact that several books have been written on the subject, that one of which explicitly adopts a digital approach to focus on American postcards. So I wanted to take a more comprehensive view. So the next thing I did was to gather all of the data from the Postmaster's General Report between 1873 and 1930. This evolved going through every annual Postmaster's General Report, finding the number of postcards at sites, entering it on air table, which was also able to tabulate this off and kind of quite fragmented data for me, and then visualising it on tableau. And what this relatively simple chart showed was really, really interesting and importantly not readily apparent from just looking at the data. What I expected to see was a huge rise starting in 1906, followed by a complete crash around 1913, I guess kind of bell curve. But actually what we see is this image here. And this is really significant because there's no boom and bust here. Instead, what we're looking at is a really slow, gradual incline interrupted only after World War I in 1917, when the price of mailing a postcard was temporarily doubled. But the number reverts to the previous trajectory almost immediately and continues along that line well past 1913 to the 1950s. So it was this visualisation that brought my attention to the fact that there's a discrepancy between the accepted narrative of boom and bust that we know occurred because of the overwhelming number of primary resources that attest to it, and the numbers that are being cited in the scholarship would show nothing of the sort, and that I needed to explain this. What I eventually realised was the Postmaster's General Report, which the Miller's citing was only referencing one subset of postcards. In its most simple terms, it only showed government issued postcards which were known as postal cards. And the explanation for this is that these postal cards were the only kind of postcard available prior to 1898 when the government had, the United States government had a monopoly on the medium which they relinquished then and allowed privately issued postcards to enter into the market. But there's very little data for these cards because in part the majority of them produced in Germany, but they were also produced in both in Germany and the United States by numerous private manufacturers. And the reason that there's such specific data for government issued postal cards is because they contained a preprinted stamp which is essentially kind of cash and therefore had to be accounted for quite carefully. So this strongly suggested that the real scale of the craze was far greater than previously had been understood that extended far beyond government issued postal cards and included these privately produced postcards, but it raised the problem of accounting for this set of private postcards and I started to search around for what kind of available data could we chart that might help us to approximate the scale of the craze. And after searching around for a while, I eventually turned to the sales of one cent stamps. I noticed that in the primary sources, stamps are identified as a proxy for the craze. So you might see an interview at a post office cashier in Atlantic City or somewhere who says that the sale of one cent stamps has gone through the roof thanks to the craze and you see this mentioned many times in the primary sources. And the data for the stamps was also included in the same postmasters general report because all postage stamps like postal cards were issued by the state. Unfortunately, the specific data drops out from the report after 1913 after a new postmaster general's report is appointed by Woodrow Wilson, but it remains a really solid and reliable data set. And so based on my expectations, again, what I was expecting to see here was a sharp rise in around 1906 and a peak in 1910. And that was exactly what this visualisation showed. So what we're looking at here and the orange line at the bottom shows the slow growth of government issued postal cards over time. And this orange line at the bottom was previously assumed to represent the numbers for the entire craze. The red line shows the expected sale of one cent stamps during this period. This is an algorithm that we've been working on that attempts to project the number of postcards that we, excuse me, the number of one cent stamps that we would expect to be sold during this period. It's still something that we have to refine slightly to take into account for rural free delivery and things like that, but generally speaking, it's relatively accurate line. And the blue line shows the actual sale of one cent stamps during the postcard craze. And we see an uptick immediately after 1898. And here we have the beginnings of an outline at least of a true kind of boom and bust cycle. So in short, the primary source upon which all this literature lied at the dates approximately right, but the numbers completely wrong. And what this rudimentary digital visualisation showed was that the number of postcards sent through the mail at the peak of the craze, which was assumed to be about one billion, which I think is already an extremely high number, was in fact closer to three and a half billion. And this is really significant in terms of the actual scale of the craze and the breadth of its impact. But, you know, obviously, you're probably thinking, yeah, but what about all the postcards that were produced but weren't sent through the mail? For example, there was a huge culture of collecting and scrapbooking postcards. So I was also presented with the problem of approximating this number, and this is where the data is the patchiest. However, there are a number of kind of corroborating facts to back up the small number of data points that I have identified, but this part is really still a work in progress. So I started looking for government tariff schedules online at the Library of Congress and discovered that after 1908 and the passage of the Payne Aldrich, tariff acts postcards are classified individually within them. I haven't yet been able to identify the specific government report that details the classification of the tariffs, but I found references to these elsewhere. And that's the data that I visualized in the final chart for the years between 1906 and 1909 based on the figures for these imports. And what I want to draw your attention to here is the bright yellow line at the top, which represents an approximation of the total number of cards imported to the US from Germany during this period. And you can see how these numbers dwarf the original estimates. So the final total for the peak of the craze represents the 1 billion postcards that are initially considered to constitute the entire phenomenon, plus up to as many as potentially 9 billion apported from Germany, of which approximately 2.5 billion were mailed. And of course this 9 billion figure doesn't even include postcard produced within the United States, of which there were many especially after 1908. So the digital tools and resources help to both identify and demonstrate the fact that we were dealing with a phenomenon that was almost an order of magnitude greater than previously had been assumed. And in terms of the significance from my research, we're now looking at a medium that penetrated everyday life in a way that we had kind of never anticipated. But from a broader perspective, what I want to argue is that this points to the usefulness of existing accumulations of mass data in digital archives that can be accessed with relative ease. And that the fact of this availability, combined with the usability and the accessibility of tools such as Airtable and Tableau help to establish an argument for the digital humanities as a kind of fundamental instrument of best practice. That is not as a supplement so much, but something more basic to the discipline of art history. You know, instead of suggesting a digital approach as a potentially usable methodology, I feel as if it raises the question as to whether we should be asking why is such an approach absent of what precludes any given project from at least a rudimentary digital analysis using sources and tools that were previously unavailable, but are now far more widely accessible. And then finally, at the same time, there's potential for a larger project here I feel I think it points towards opportunities to engage with other digitised collections. I'm thinking of the New York Public Library in particular, where you might be able to take advantage of either a digital content analysis, and or indeed the emerging technology around a handwriting analysis, or perhaps most interestingly, a relational analysis of these two facets. Thank you. Fantastic. And yes, there are all sorts of applause, visual applause, visual applause around the room and beyond. Thank you. Thank you for that. Fantastic. And thank you for some mind boggling numbers in there as well. We've sort of moved in terms of volume to the billions already, and it's not even one o'clock so fantastic for that. I will keep moving as I say, do keep up your if you've got comments about the thing on the tech side or, or about the event as a whole do person in the chat questions for the panellists to have ready, put them in the Q&A, they can be lined up for our discussion section. And that will move straight on now to our third, third and final presentation in this morning panel, the panel one. And we're sticking with Duke, with Hannah, Paul and Lee. Hannah L Jacobs is digital humanities specialist for Duke University's digital art history and visual culture lab, visual culture research lab, and is studying for masters in information science, the University of North Carolina at Chapel Hill. Hannah L Jacobs is a professor of art history and German studies at Duke, and is co director of the digital art history and visual culture research lab. And Lee Sorenson is a librarian and bibliographer at Duke University and another co founder and co editor of the dictionary of art historians. And it is indeed the dictionary of historians studying art history at scale, which is the title of this presentation so welcome our three presenters over to you. And thank you so much for for having us. I'm going to kick things off and then I'll pass things over to Lee. So the dictionary of art historians is an open online resource that documents the lives of art historians of mostly Western art. The project is supported by Duke University's digital art history and visual culture research lab with Lee and I as co PS and Paul is our research collaborator. Through the lab and funds from Duke's office of undergraduate research support we work with student research assistants to draft new entries, revise existing entries and continue to improve the data structure. And this presentation will be sharing the project's history, showing the latest version of the website and database and discussing possible research questions that the date that the dictionaries data might support. Thank you, Hannah. Like few contemporary database sites, the dictionary of art historians was begun as a paper project over 40 years ago, and has developed as the internet developed. It's a specific mission for graduate students to research significant Western art historians by their methodology. It includes only art historians who have died or at the end of their careers. There are over 3000 entries covering globally historians of Western art. Each entry focuses on a bio bibliographic form, the art historians familial background schools they attended and faculty under whom they studied. Art historians writings in the chronology of their lives, as in fact they happened. It identifies debates and animosities concluding with bibliography of their own writings and sources of the information used, and the art historians archives, if available. It is similar to many contemporary databases, the dictionaries goals and structures broadened and sharpened by the examining crowd use and feedback. When the project joined Dukes digital art history and visual cultural research lab in 2017. It was brought up to date functionally by loading the static files into a content management tool and migrating it to the Drupal open source software. It is a collection of the processes of our second presenter and the projects co editor Hannah Jacobs. The previous clunky way of keyword searching was replaced with the construction of discrete fields for the most common queries gender institution country of birth and death, and subject specialty of the historian. It is content. We were lucky to be able to use the early general art historiographies, which were the beginning, which are beginning to appear. In addition, the published collections of specific groups of art historians appeared. We were able to mirror current scholarship. One of these was the massive inventory of German art historians who fled Nazi Germany between 1933 and 45. It is important to note that Paul will speak about. It's important to note that the data of when lans book is still only available in searchable form in our dictionary. Today, the project boasts of unique data groups of art historians, most notably women historians and those of color. Representative data visualizing and graphically analyzing our discoveries for the first time. We are truly able to study art history at scale. Thank you Lee, and we're going to take a brief tour of the latest version of the website and database. The dictionary's website and database are designed as a tool to help researchers working at a range of scales. So from studying an individual as Lee mentioned to investigating themes across many entries. The site is built in Drupal also as Lee mentioned and this version that you're seeing today has just been launched in Drupal nine and we're going to be migrating it to Drupal 10 very shortly. This is a major upgrade for the site and not just in terms of the platform but also in terms of some of the data work that we've been doing to improve scalability. Individual entries and provide basic demographic information such as names birth and death dates birth and death places countries of origin or residents and gender. And we also document institutions that art historians have worked at the types of careers they have had and their major fields of study. And Lee did not mention the careers I'll say because that is one of the newest fields that that we've just made available. As mentioned our research assistants work with Lee to draft biographical entries the overview field in which we also link between entries to help readers experience the many relationships that are represented in the dictionary. Contributors are also able to develop a selected bibliography of the art historians work, a list of sources used to create the entry and when available links to archives that hold the art historians papers. Each entry is attributed to all those who contribute that entry, which allows our students to point to their work on resumes and applications. On an individual entry page under more resources researchers can access links to other prominent data repositories and online collections that contain relative relevant information about an art historian. This feature developed out of a 2019 workshop with the Getty Research Institute, as well as an earlier partnership with the Getty Research Portal. And you'll see that the more resources links include the archives of American art, the dictionary of French art historians, the digital public library of America, as well as the Getty Research Portal and the social networks and archival context project and wiki data and WorldCat. To help researchers find entries of interest, we continue to offer a full text database search. But for those who want to see what's represented in the dictionary, we have a browse page where they can search through entries alphabetically. Folks can also filter entries by gender country subject area institution and career type. We also provide a directory that lists all of the entries by name along with some key contextual data as an as an alternative form of browsing. Entries data fields rely on a number of data standards, which are documented in the about section of the website under data access. Entries data fields are mapped to the Dublin Core metadata standard and our standardized vocabularies include the Getty's thesaurus of geographic names for place names, the Getty's art and architecture thesaurus and homosaurus. We also add our own localized terminology where necessary, and this information is documented in our data dictionary. Scholars wishing to study many entries, we offer an API that allows researchers to download data or to connect our data set with their own in a separate web application. All content is licensed under a creative commons attribute share a like license to encourage digital digital scholarship, and the data are currently available as JSON, but we can make a format such as XML and CSV available as well. And we're also in the process of submitting the data set to Duke's digital repository as an archival record of our work to date. And just to give you an initial sense of the makeup of our data set, here's a visualization showing countries of residence for art historians who were active in the 20th century. Likewise, a visualization showing subject areas represented among 20th century art historians. These two visualizations were created by my first exporting the data and doing some minor data transformation. And then I shared that data with our labs graduate assistant data Hogan, who was able to plug the data into the visualization software tableau to produce these exploratory visualizations. Finally, we have an extensive and multilingual about section that describes the project, a blog where we post periodic updates about our work and our contact information, and we always welcome information about art historians who should be included updates or corrections to existing entries and other expressions of interest. So now I'm going to hand things over to Paul, who will share some of our digital research that draws on the dictionaries data. Thanks Hannah thanks Lee. Yeah, there we go thank you. Having looked at the origin and structure of the dictionary of art historians I want to turn our attention to how exactly one category of evidence mentioned by Lee can extend from the DOAH data to address a significant question of art historical research. That focuses on well known aspect of European Euro American art history, namely the migration and exile of Jewish and leftist art historians from Nazi Germany 1933 to 45. Such a history is well known in terms of individual biographies of course like proud timer, but scaling up the question to include the large number of art historians and the DOAH, who were exiles allows us to see patterns and patterns that raise new questions about the history of the spread of art history itself. In addition, this extended analysis also allows us to reflect back on the other sections, since dealing with a large number of art historians in aggregate points back to the data structure of the DOAH in the first place, as well as how the site may be extended through linked open data, and attending to interoperable functionalities. Having looked at the structure of the DOAH, as well as visualization possibilities, how can we test its potential with the subgroup of exiles. Here we start with the source that Lee has mentioned and contributed much to the DOAH, Orica Wendlands Biographies Handbook, We developed data set from these art historians DOAH and then extended that with additional geographic and temporal data from Wendland. From this we were able to map the movement of art historians from Germany to the world, such as this map of the final post war destination of all German exiles listed in the dictionary. And the other maps we have created show known patterns like the early move to Paris in London, for example, but also more unexpected locations, such as the importance of non coastal locations in the United States. Furthermore, an animation helps us to see the clustering here in the United States of the initial move in exile, but also the move within the United States of exiles, possibly indicating the precarious nature of their unemployment as they moved from site to site. The point here would be that the biographical nature of the dictionary has led us to non biographical systemic categories that then form a context in turn for the distinct entries of the DOAH. Working at different scales pushes the DOAH to the level of a research site and also expands its potential. In this regard, we are thinking about how we can start to connect systemic questions with our biographical dictionary. For example, we have begun to extract questions about social and intellectual networks by pulling out bibliographic information of the DOAH, such as this map of the exile path of any art historian who was active at the Wabog Institute, both in its Hamburg and London iterations. I would argue that this map is in and of itself not that interesting, except as an exploratory site. The visualization in this sense may lead us to other questions, such as whether these specific exiles also shared professional or intellectual trajectories after leaving Nazi Germany. For example, see patterns in terms of which journals the Wabog group of exiles published are these patterns indications of standard career choices and publishing venues like the art bulletin, or do they cluster around editorial boards through the social network of the institution, or content from intellectual communities. In this case, the extension of the exile work from the DOAH leads us back to ask systemic questions of the DOAH as a means of expanding its scholarly use. In the end, producing and using the DOAH allows us to think about the intellectual questions and preconceptions that were specific to each phase of our analysis. As we thought through the structure and history of the DOAH, as much as we learned through exploring its digital and art historical possibilities. We also learned through the process of entering and visualizing data, a process of knowledge building that is certainly well known to the digital humanities. Ultimately, our goal is to see scholarship, in this case art historical, and resource building as dialogically engaged intellectual practices that are too often separated institutionally and epistemologically in our worlds. And I want to emphasize that last point that resource building and scholarship are integrated activities, not necessarily distinct processes. Our argument calls for more engagement between the areas of digital and evidential expertise in order to draw out the complexity of critical questions concerning human experience, as well as its current expression through digital and art historical means. Thank you. Thank you so much and we're looking forward to the Q&A here again as our website. We encourage you to check it out. Thank you. Thank you all again, sort of round of applause for sharing some of the kind of front of house and back of house of such an amazing resource. We've got a couple of questions lined up already, which we will come to in a moment, but if you've got comments, observations on any of the papers, any of the presentations this morning, do post them in the Q&A. We've got a bit of time now for discussion around the panel, so if everybody, everybody who's speaking this morning should, I think, kind of unmute and be ready to jump in. I mean, I've got a couple of thoughts really, but I've got a very kind of basic question, a probably rather stupid question, which you may want to dismiss immediately, which is about, well, in all three presentations, there's a lot of emphasis and importance put on visualisation and the possibilities of visualising data, visualising information, not just because it communicates effectively or does a good job or is instantaneous, but because it can actually reveal things that you would not be aware of otherwise. There's a kind of revelatory dimension to visualisation. With that in mind, I wonder, perhaps around the panel, what attention you give, if any, to almost the aesthetics of that visualisation, even things like the choice of colours. I mean, it's very striking, you know, anybody who has to kind of make an Excel chart, you know, when the column that comes up pink sort of jumps out and the column that's pale blue recedes. I mean, I wonder in terms of your own projects and your own work, what emphasis do you put on those aesthetics, how much you depend on kind of off the shelf kit in order to a pallet which is just ready or if you've had experience of working with designers and whether that's part of the process as well. I'll make a comment. Yeah. What you tried to do with the design of visuals is decrease the cognitive load on people, the amount of work they have to do to interpret what it is. So that's why you use align colours and align with common uses and organise things that, you know, in a timeline format consistent with what people expect because you're leveraging what they already know. One broader way to think about this that I've found useful is that what you're engaging is conceptual toolmaking. Whatever you're doing with the data, the actual interface with the yourself and other people is this interface and that's the conceptual tool, the effectiveness of that tool is really important. Jane, what about your visualisations? Is that what's the working process there? Are you kind of taking things off the shelf and working with that more? I'll just jump in quickly there. I think that one of the big issues here is that we're working with pre-packaged software and that that itself comes with its own expectations and at least in our lab, we talk a lot about that and we talk about the idea that, well, you know, if you're using Tableau, that's really being used by the advertising industry and that has an issue, you know that there's a meaning to that. One thing that is really interesting for example just in the dictionary is of course that, you know, Hannah and Lee have spent a lot of time thinking about the user end but, you know, the maps we're producing in, you know, we're just using ArcGIS Esri, you know, we're trying to make them look very drafty. We're not trying to actually make them look very good. And we call them draft maps and the reason is to kind of point to that a exploratory function but also point to the weakness of the methods and the tools we're using. You know, the fact that it is about a kind of connection to a privatised economy that is different than what we might prefer. So I do think that actually that is a really interesting question of how we think about visualisation and that's intersecting with a kind of broader digital economy. And do you see kind of futures being mapped out there? I'm thinking about kind of animation and we've sort of quite a lot of static images that I think there were references to some of these things being digitised and kind of having a duration as well as a kind of slide like quality. Is that something which is developing in your own areas or which you're utilising yourselves? I'll note to you the area of digital twins which is emerging in the corporate areas where we're trying to map a lot of different types of data onto, for example, a model of a business or a model of a race car. And the purpose of doing that is then people are aligned in their models and the data all fits together in a meaningful way that relates to the real world. So that's a particularly powerful approach. Picking up on what was said before, in the initial stages where you're exploring things, it's not terribly useful to refine things too much. You're just hunting for ideas and perhaps it's at the presentation where you put a fair bit more work into it that you'd be a bit more elaborate. Yeah, I would just add to that, just as significant as the visual effect, the visual impact of producing these data sets and the kind of whatever forms we're producing them in, in my instance, its graphs, is the opportunity, which is I'm working on at the moment, which is to make them interactive to some extent. I think that part of the force of the visualisation can be, or the force of visualisation can be harnessed and enhanced through allowing people to interact with it themselves and to truly tease out the significance of the comparison between and amongst different aspects of the chart. OK, great. I've got some questions lined up here, so we should start to move to them. I've got a question posted by Lena Kraus, and it's for the dictionary of art historians team. I'm very interested in the API and the reference to other authorities such as Wicked Data. I tried the links on a few profiles, but they bring me to an exemplary query with Georgia Fasari, and I can't see the authority IDs in the JSON file. I was wondering if you have mapped each person in your dictionary with the other authorities. I imagine it would include work to contribute to Wicked Data, since some art historians might not have a Wicked Data node yet for the example. Are you interested in doing so? I can take that question. Excellent question. The way that we currently have it set up is more in a one-way system where folks who are looking at an entry on the dictionary can do an automated search, so those links are automated searches of the other resources. I will say I've noticed there's a bug, so if you click on one and it's pulling up an example entry or it hasn't searched for the correct person, there's something that I'm working on there right now. So you've hit on something that is a challenge with the recent launch, but the intention is that it will search for whichever art historian the page that you're currently on. As far as including IDs and connecting into Wicked Data, that is an aspiration for the dictionary to actually have some of that data pulled in and directly available in the API. It's not currently there and we do anticipate that we would be contributing to some of these resources. Some of the resources are limited. For example, the Archives of American Art is not necessarily going to have our European colleagues listed, but for things like Wicked Data and perhaps WorldCat there may be opportunities to make suggestions. That was a core part of our partnership with the Getty research portal actually was as they were identifying texts that they should digitize for the portal. We were in conversation with them and sharing data with them so that they could identify authors to include. So that's something that we're hoping to work on in the future. Okay, great, thank you. Sarah's got a hand up. An actual hand. This isn't really a question because I'm just trying to shape some ideas, but I'm just picking up on an idea that I think was presented by Paul right at the beginning. The idea that doing this kind of work has the potential to focus or shine light on things that have been missed perhaps or left out of the grander narratives or conventional traditional forms of history writing, let's say. It's probably kind of paraphrasing and reducing your arguments a lot here, but it just made me think about, you know, and I think maybe we'll explore this in some of the later presentations as well. About whether mass data methodologies part of the impetus or one of the very strong strands of this research is work is kind of shifting the kind of conventional gaze of our history from the object or questions of aesthetics and normal questions to just the things we've heard today about, well, people, institutions and networks, you know, sort of postal services, it's sort of shining a light on, you know, just some of the uncanonised of history. Again, sorry, it's not a question, but it's just the thought about, you know, does mass data methodology have that potential to kind of yet on earth things that have been left out of conventional disciplinary structure is where I'm going with that. I'll make a brief comment. Yes, I think it's really important that it does. There's a broader societal function to creativity and art and the increasing print towards technology and shifts in the careers that people are taking are going to challenge that in various ways. And you see that too with artificial intelligence, with the creating artworks or music works or authoring things. We need to understand and find ways for people to participate and understand their own development in an artistic sense, quite apart from what the canon side of things is all about. And yes, it's very considerable potential for that. I think as well as an eater as well what you was, you know, the history of the Royal Academy isn't really written about exhibition length. You know, it's about Joshua Reynolds and, you know, the sort of formation of national schools, you know, that's that's what's tended to come to the fore, but I'm kind of interested in the kind of intersection of the kind of. Yeah, then both bureaucratic, like the number of days, the sales, you know, and actually the intersection of those with then writing about careers and aesthetic choices of artists. So I wonder, Anita, if you want to just pick up on that as well. Well, and also just by dropping that data into different visualisations, it's then the possibility and just by looking at those trends using your eye to say where are the outliers? Where's, where is anomalies? And then to go back into the data and say, well, did this occur because there was a, the Royal Society made a change about who could exhibit? Was there something global in terms of, of course, you know, the non-academicians that were less during, you know, wars? And if that's the case, maybe then the Royal Academy is engaged in broader rather than just looking at it within London and within the London. Kind of ecosystem of art history. And to Paul's point, I think the possibility of creativity with mass data to allow the public to interact with the data, I think is critical. And I think Shane sort of referenced that. The other thing I, the tool that I actually built for my PhD use Neo4j and it, and it visualises relationships between framers, artists and frame makers. And that uncovered links that would have taken me years to find with my bits of paper everywhere. Like that was phenomenal in terms of visualising relationships. And then I would go back using traditional historical methodologies to follow up those links. I mean, it's striking how already today there's been several points made about the need to keep different, both the kind of the established methodologies, the practical, all the kind of analog methodologies in dialogue with the technology as well. I mean, several of you have said that already. There was a question that was posted earlier on that connects with this discussion a bit, but I see that Ming and Hans have hands up as well. So Ming, I think you were first. Thank you, Martin, and thank you everyone for those fascinating papers. This ties in a little bit to what we'll be speaking in about in our paper as well. So I won't sort of jump the gun here, but I'm wondering to what extent all of your work, and I'm addressing all of the speakers here, thinks about the question of bias in data structures and the bias that is embedded itself in the data, which, you know, as we know data is not a neutral category. And so how do you contend with those issues in your work? I can jump in with one sort of obvious example. When we're covering women or art historians of color, they're not only are they documented last but when they are documented they have shorter entries and so we're in kind of a predicament. How do we fill in a lot of places that are just missing from those people. And it's a concern. In some cases we've been able to use our own knowledge or whatever to fill those things in, but in a lot of very precise visualization, we do understand that there's some bias there, that there's absence bias in the data. This actually does this does link to the question that I was kind of that was posted and I can't. I think it's kind of got deleted now, but there's a very simple question. I'm kind of perhaps a kind of deceptively simple question for the depiction of art historians which was about, do you do you cover independent art historians as well as people with institution affiliations, but out of that you can sort of gloss that I think well how do you define an art historian I mean how are you creating your data set in the first instance. We come across this question a lot, and actually we have a flexible definition for groups that were excluded from the Academy in many cases, we look at their effect on the public their profile. Generally we say if you published something that was art historical, you fit the first definition of what an art historian is and for people who were again we're prevented from from doing much more on that that's good enough for us. And the dictionary, I should say that the the selection of art historians is not our own it's not our favorite art historians. We began with the four major bibliographies coltormon and and Bethan and things like that and then as we wrote those entries, the people who were mentioned built it in that way so, while we have adjusted for the the minority fields, and basically itself generating. Yeah, because it's so interesting here also with Shane where you're changing the criteria of how you create your data set can completely transform the overall historical picture. I understood what you was that you're saying. I'm correctly. There's hands is going to stand up and then react. I just want to add one small thing to that Martin. One small thing and it goes back to your your reference to white and white. You know this this is not a new art historical problem. If you look at Foufflien's principles of art history, you will find a shocking number of paintings from the Alta Pino Catech in Munich, why because that's walking down the street what he could see. So that was his data set, or at least obviously he was working more I'm exaggerating but so this is not a new problem that we have to deal with. And I think it's always going back to Shane's comments it's about well, this is now a fundamental part of our method just like knowing which, which access we have to which collections. And I think I think what I was going to ask is very much pointing into the same direction because I was also struck by Shane's approach where, you know, drawing on completely different data sets changes the parameters substantially. And that's exactly what I was to saw as an interesting question for the dictionary. Because. Well, there are many art historians who literally have left probably no paper trail British art historians for example, who even taught at universities, but who have for example you see a criterion Lee, who have never published as an art historian, simply because that was the way that you know academia was structured at the time. I find it very interesting there to think about the frameworks that we impose on our work by the by something as simple as saying we're doing a dictionary, which by definition is of course based around prosop prosop graph prosop graphic approach right that we that we do look at individuals where we have at least a certain amount of data. And that's because you could of course also source thinking about mass data, all the names from course syllabi or even student lists. People who in depending on your definition also count as art historians, but as I encounter in my work quite quite regularly are virtually, you know, not documented in other in any other meaningful way. I'm really interested in how the kind of the type of the type of publication that we're tackling, for example dictionary also imposes a certain. Yeah, selective approach as to what data we mine primarily, and it's very interesting, I think to think about how to push the boundaries or something like a dictionary towards people that are literally undocumented, which do you have a lot actually of artists, you know, where it's just say a name, and then work around 1739 or something like that. And to find an art story in certain parameters, I mean to finding an artist is completely fraught almost by definition. Is it about self identification is about exhibiting is about making a living as an artist. Well, you know, Rhian, you have your hand up as well. I was just going to go back to the main question. And basically, my data set anyway, and I'll go into this in more detail about multiple scale methodologies. And my data set is based on stratified something cross multiple institutions and resources for the exact reason that I want to mitigate previous problems of the data collection and with institutional record looping. But I have created an extensive appendix, I don't expect my examiners to read so that if anyone wants to use the data set in future and I really hope they do, they can go back and critique expand on it. My, my data set is representative is certainly not exhaustive. And I think that's, I think it's really important that we bear in mind that whilst we're talking about these projects, and that they can keep, we can keep getting more resources, sorry, more information from them. That actually they can all start into linking and we can start seeing things from different perspectives, but only if we question how that date has been gathered in the first place. There are a couple of questions in the Q&A, one of them I can see Sarah Turner is answering live it looks like, which is a kind of technical question about the Chronicle 250. But there's a question again for the dictionary about historians, but I guess it does kind of connect in with this issue about classification and categorisation is from Laurel Zuckerman. Does the dictionary distinguished between Jewish and non Jewish exiles from Nazi Germany, for instance categories for Holocaust victims or Nazi party members. That's a great question, but I'll see. Yeah, and I think it does we don't have field for religion, because it's tricky in many cases, but if you're interested in this you can do a keyword search with Vennlon and je w which gets Jew Jewish, whatever like that. That should be a pretty discreet list, at least from Vennlon of who the Jewish Holocaust victims were and if you don't get that and then the Vennlon one. Those would probably be people who were of other either religions or leftist as Paul said. That's a good example of how the dictionaries itself of kind of jumping off point so we went back to Vennlon and we pulled out Jewish non Jewish, in order to make that distinction in our maps, which then can lead back to questions of the actual say Jewish art historians who are in then the dictionary itself so it really is a two way street there. Yeah, it leads you into tension kind of fraught questions about identification self identification also struck, you know, you've got to get a fairly binary gender classification system and play and of course that's there can be questions about that in terms of perspectives on identity and self identification and self understanding. I shouldn't say that the dictionary has multiple identification fields it's not just binary. Well, I think it was male female or other wasn't it, or male female or not known, I thought was all the. But that's that's what that is to say that there's a whole range of ways of searching. Okay, do we have, I think something else can just come in from Sally Woodcock. Apologies for a rather nuts and bolts question. So, in this context, if, if, if any context that you need to apologise for nuts and bolts question in the early 1990s. I used a relational database called idealist which for its time had unusually sophisticated search mechanisms, but also turned out to have dreadful data exporting importing. It is long obsolete, but I still readily use it for a data set of about 9000 artists. I've been trying to find a replacement that is available offline picture stores where I sometimes work usually have no internet access. I did look at air table mentioned by the speakers, but it appears to be online only. Can anyone suggest a database that is capable of holding large text fields and working with dates and numbers which we doesn't always need an internet connection. Many thanks. It's interesting. I'm going to make some point as well about how dependent we are on online resources as well now. But Rian, you have your hands straight up as if you may have an answer or response. Yeah, well, if my, I used Microsoft access for mine, it does require you obviously importing all the data and the relationships. But it was quite straightforward. I mean, it basically I do use mine online because I use my remotely through my university because I use a Mac and access isn't available on Mac. But that worked pretty well for me because I had a similar concern that if I was too reliant on the internet, would I would I be caught out. And also, would I ever be able to download the scale of the data set. This is before I knew about all the tech available. And it was also free for me to access. So yeah, I'd recommend it does have it. Sometimes it has its limitations, but if once you know, get familiar with it, it's great. And my university anyway provided a whole handbook on how to build my data set and gave me training sessions and stuff. So I always want you to pass on that information if you need it. I'll hand it as well. I think about your hand. Yeah. Yeah, that's, I think that's all excellent. Just to speak to an alternative to air table there's an open source project called base row and I'll put the link in the chat that has created something that is very similar to air table and because it is open source you can choose whether or not to run it online or actually just host it on your own computer so if folks are interested in alternatives to air table that is there. After that comes over comes a more more general question for the for the panel, particularly the panel who have panel members who are primarily art historians or historians by training which is how do you get the technical know how to produce to generate this these data and these visualizations it. I mean, is that is that access to training. Do you just kind of find your own way. Through a university, they have a remarkable amount of training available for PhD students. And also, I think it's about showing initiative to walk across campus to, you know, the computational areas and it actually allows you to take different subjects so for example you can take. I actually think it should be compulsory for art historians in first year to be able to take undertake some kind of computational software learning. The challenge that we have in our art historical discipline is that the professors look at something and press print, and you will have been into many a professorial office which is full of books and lots of drawers with files in it and so that older generation is. And the, and the idea of the commiser is not linked to using digital technologies as part of the research. But yes, within an educational institution there's a remarkable amount available. There's also quite a lot online if you're proactive. I was going to say Duke University Libraries has a data visualization lab that anybody can come in and play with the software they don't do the projects for you but they have experts there who can suggest software and show you how to use it. Yeah, just I was similar that I had to do a lot of cold calling. It was exhausting in the first year because I had my idea for my methodology and I didn't want to compromise my methodology by not being able to access the resources and again I mentioned this in my talk about the fact that I didn't have funding to build software. I had to use everything that's free and available remotely. Sorry, my dog wants to say hello. And so yeah, I had to physically, in some cases, go to other members of staff in other departments and beg for their time and obviously they're stressed in their own capacities. So I went to geography, for example, for a 3IS support. But also, I went to other institutions like British Library, Biggie Labs, and again just begged for help. I was not one of their funded students. So I think it would be really interesting later to consider what resources we can pull or pull together from these discussions to ensure that other students don't have to be in such awkward positions because I found it quite intimidating to go into a sort who is new to these institutions to have to go and beg, borrow and steal for these resources. But then once people got through the door, people were more the way to help. Great. And Shane, over to you, and then there's a question lined up for you as well. Yeah, I just wanted to add briefly, at its most basic level, obviously engaging with these methodologies is no different to engaging in any kind of new area of exploration, but I agree with what you're saying about it being more intimidating. And I think that what Paul Wilson was saying about this idea of starting small and doing it incrementally and taking it step by step is using the, as Rhianne said, the kind of resources that are initially available to you online. I think this is a really effective approach to getting to the more complex things because when you see these mass visualizations initially, there is a certain amount of intimidation about the menu. It is, you do struggle to conceive of how you get from where you are right now to do much more complex kind of visualization. And but I think it's the same kind of slow intellectual building that we do when we investigate any new area, just that it's most basic level. So, yeah, you have your hand back and I've got a question for Shane who might come back there. Yeah, absolutely. So I'm not an art historian. And I will just say that one thing I would emphasize here too is collaboration and seeking out partners and that can look like the kinds of consultations that have already been described right universities that have those kinds of resources available. There are also plenty of workshops and institutes that are held around the world, some online some in person, and some that have funding schemes that are great, both in terms of of gaining skills but also in terms of building community and looking for for potential collaborators and that that's something that I highly encourage folks to do. Okay. There's a question posted here for Shane. I think looking at the time this made for us to close for this this session but is posted by Lena Kraus and question for Shane. Could you tell us more about your methodology to gather your data. Did you automate some of the extraction from the digitised reports. That's an excellent question. Unfortunately, the OCR on the digitised reports and the nature of the way that they're laid out meant it was kind of not possible to extract them in any kind of efficient way. I had to go through them each individually and pulled them out. And a second impediment to that was that every time a new postmaster general is appointed, they seem to decide to reformat the report so, and that's part of the reason why some of the data drops out in 1913. There was nothing sufficiently standardised about it to enable me to even start thinking about how I might do it at scale or using anything other than the most rudimentary approach for that, fortunately. But thank you for that. And I think it nicely illustrates how already today we're getting that sense of how these kind of details of the archive, the detail of methodology can actually open up and kind of can transform your methodology in quite dramatic ways. That as well as that dialogue between between different methods kind of established methods and new methods which are facilitated by technology. And we are running up to our break, which I think is that kind of, it's not billed as a lunch break, because we're not providing you with lunch, but it's that time of day, I think. And it's been such a fantastically rich and stimulating, stimulating session already with such a range of approaches and questions arising out of it. So the conversation is going to continue with this afternoon's panel and then with Hans contribution towards the end as well, which is going to I think offer some more kind of historiographical reflections and conversation coming out of that. So, for now, I think you could join me in thanking through our virtual applause and various expressions and gestures. I'm thanking all our panellists this morning Anita and Paul, Shane, Hannah, Paul and Lee. Thank you so much. And we will return at 215. But for now, round of applause and do check the chat as well. There are various notes and suggestions and links posted in the chat for your reference as well, but otherwise do enjoy your 30 minute break or so and we will see you back in the room at 215. I've sort of half joked before I'm sure that sure there must be some way of getting sandwiches delivered to people's homes during breaks on online sessions, but we have to do a menu of options, and we feel like breakfast options for the US. Of course. Yeah, time zones. Australia, I know a need for Anita and Paul, I think it's some pretty unearthly hour. Anita, what time are you on Tasmania? It's quarter past one in the morning. Oh my goodness. I think that's like Horlicks or Martini, depending on your body. Hold the sandwich, send the, at this stage, probably sugar. Yeah, really chewy black coffee. That's great. No, I have to say things like this is certainly one of the silver linings from Ovid. Oh, I don't be able to connect up around the globe like this. Yeah. Different world. Next time you have one of these events, have Anita show you some of her other words. She's done some really amazing things with Neo4j. Well worth a look. That's collaboration in action, isn't it? But do keep posting. I see there's quite a lot posted in the chat. I think people do take note, and they really appreciate chance to see any links, anything that you've posted, anything that you put up. Anyway, people coming back in. Well, I think I'm noticing it's 215. So, and it looks as though we have most of us back from our break. I hope everyone had a good, a good kind of refreshing, tasty break and lunch, and it's nice to see you all again. So we're kicking off our second panel in the mass data methodologies workshop today. My name is Bailey Cards and I work at the Paul Mellon Centre as an editor of open access online publications and some of those have been mentioned already such as our journal British Art Studies and as Sarah said we really see those publications as kind of testing grounds for exploring how work using these kinds of methodologies can can be shared and communicated so it's been an amazing workshop so far and I'm really looking forward to the series of papers that we have coming up. I've been told that I don't need to run through the kind of zoom housekeeping again, which is great but just for anybody listening just to be aware that you can share questions at any time through the Q&A button. You also can share messages in the chat and those will be visible to everyone and you can just kind of comment on that at any point. And you can also raise your hand if you would like to be called on to ask a question for yourself. I'll dive kind of right into our first paper and introduce our first speaker. So we have Rian Addison McCreiner, who is an AHRC collaborative PhD students between the University of York and Tate Britain, and was formerly curator of historic fine art at the Whitworth University of Manchester. She will be sharing her paper, multi-scaler mixed methodologies, rediscovering landscape artist studios, and just to note that's a different title than we had listed on our event, which I think isn't correct, but that that is the correct title so thank you, Rian, take it away. Hi, thank you. And, sorry, I'm just re-gilling screens to make sure I can see everything. So my thesis indoor space is spout or mines, landscape artist studios 1780 to 1850 explores the irony that all landscapes were being painted in urban London. The outside world is being created inside the limits of the room. When a scholarship is considered the studio con galleries of landscape artists is being predominantly through monographic photographs focusing on less than 20 artists, replacing the wider landscape artist population and skewing our understanding of their production of landscape paintings and how we interpret them today. To have a clear understanding of the landscape artist's urban experience, I applied multi-scaler mixed methodologies starting with the landscape artist community on a macro level and metaphorically speaking zooming into the studio. I would discuss the multi-scaler methodology a little later, but first I want to focus on the mass data aspect, which formed the foundation for my research. So why use mass data methodologies that some people have already touched on it already. As Martin Myrone and Diana Greenwald have each discussed, mass data sets will consist of art or artists which are not perceived as masterpieces or did not attain celebrity status, having their presence provide a wider social context, and thus are historical context. Arguably the greatest social context that these sets provide is a seemingly more democratic, non-hierarchial overview of artists' populations. Paris Spice Gans has argued that this requires a willingness to use qualitative data and qualitative data approaches to address the idea that a canon is flawed when it fails to truthfully represent the period it aims to embody. Thus, in order to have a more accurate understanding of the landscape community, I needed to create a database which set aside previous scholarship, step back to look at the breadth of the community, and return to primary evidence where possible. For a macro perspective of the landscape artist population, I built a database and GIS map, which you can see a brief overview on the screen here. The database holds details of around 2,700 landscape artists at around 6,550 addresses. This data goes well beyond the numbers that current scholarship implies, not only establishing where the studios were located, but revealed how prolific the genre was and exposed a neglected artistic community. Using stratified sampling, that is a stripped set of parameters, I created the data from exhibition catalogs and London directories, which listed artists' addresses. Using stratified sampling allowed for cross sampling of the data from multiple institutions, which could be compared to provide more comprehensive data set and clarify errors and contradictions. It also mitigated but does not eradicate institutional biases and resets the reliability and validity in the data. And it also allowed, as we've mentioned before, crossographical analysis, reconstructing the landscape artist community using accumulation of scarce data. In short, a stratified sampling ensured that whether previously overlooked are successful, professional or amateur by their own definition, each person sits equally in the database as a landscape artist. Critically embracing advanced technologies has allowed for the social interpretation of data which is not possible in previous scholarship. This included identifying clusters of addresses, the proximity of landscape artists to one another, and making queries by location to show the geographic significance of the urban studio on the creation of landscape work. So, I'm now going to show you a brief video demonstrating the database and map. I've deliberately pre-recorded this to hopefully ensure that there's no lag or anything. So, here we go. The database was built in Microsoft Access, which allows for the creation of different tables, which form relationships, as well as being able to explore the data in statistical analysis. You can see an overview of the six tables that each link according to a key field. So, for example, every artist's address and source is designated a unique ID number to prevent duplications. This means that an artist can have multiple addresses and each dress can have multiple sources. On the left panel, you can see six tables. From these tables, I can then create queries which are listed below. Queries draw on specific fields to create another table which you can then use for calculations of filters depending on your question. The table 1, artist, holds immediate information about the artist that our unique identifiers. Table 2, artist addresses, allows for artists to have multiple addresses, their unique ID number corresponding with geotagging on the map which I will come to. Table 3, address dates, allows us to track the order in which artists were at any given property and accounts for the returning to addresses on multiple occasions. The other tables 4 and 5 are a bit by institutions and suppliers and table 6, sources, does exactly what it says on the tin and was essential for clarifying contradictions in the data. It also ensured internal validity that someone else could replicate the same findings in the future. Let's look at an existing query. Here, what I'm about to bring up, are the London addresses for John Constable, which I extracted from a case study. His artist ID is the number 105, which is linked to multiple address IDs. The address dates have also been pulled through to the query, allowing me to filter in order of occupation. I went to the query further. Constable lived it on Charlotte Street, Fitzroga, twice in his career, firstly in number 63 and later in number 35. To establish his interactions with other landscape artists, I created queries to the street and for the two separate time periods. So here, for example, we can see the other landscape artists living on Charlotte Street when Constable's living at number 35. Caution must be applied however. This is where technical and historical knowledge is required to interpret the data. Although these artists are listed on Charlotte Street, there were numerous iterations of Charlotte Street throughout Fitzrovia. So if we cross-referenced these artists with the accurately mapped ones in Fitzrovia, it filters down to considerably less as we shall see in a moment. So whilst the database contains over another hundred artists on the Charlotte Street, we can't be certain that this is the correct Charlotte Street. Oh dear, I apologise, I've just messed up my own video. Start that again. I built a map using open-source software QGIS. The purpose of the map is to be able to, in the picture in my son background, the purpose of the map was to be able to process the vast number of addresses and place them in a visual context known as a meta-image. This allows for a greater understanding of the locations of landscape artist studios and their proximity to institutions, other locations, and provide real-life information such as walking times between locations. Here is an overview of all the landscape artist's addresses between 1780 and 1850. My main source for this thesis is Richard Horwood's Plan for Cities of London, created from 1792 to 1999. Horwood's map is the earliest to list house numbers in London, thus allowing the marking of the good locations to be as accurate as possible period. The metropolis expanded exponentially between 1799 and 1850, so I overlaid VR Davies' 1843 map of London to create a visual parameter of the city. I also used OpenStreetMap, so that's VR Davies in a bit more detail, and I used OpenStreetMap for a 21st century perspective. Horwood's map was georeferenced by several other people, I can't take credit for that, using surviving monuments and road junctions to align with the real-world map. Although the alignment is detailed, we must consider that multiple layers, which have been hand-drawn, will contain elements of human error or artistic licence. The maps have been digitally walked to accommodate the slight variations from the modern OpenStreetMap, so the location of the StreetMap on Horwood's map may be 50 metres west of where it is in real life, for example. On the left panel, you can also see the data files. The map works in tandem with the database. Each layer of the data points can have an attribute table attached, which in my case contains a brief amount of information which corresponds to the database. So, for example, this is all the addresses of female landscape artists, and it should bring up the attribute table now, which includes their address ID and artist ID. I then imported additional information from the database such as their names. As for the database, you can use filters to ask queries and save them on different layers. You can filter through the data or select points on the map. So, for example, here I'm going to show how I can select using a polygon around Soho Square. Or, apparently my brain's working faster than my video. So these are just the artists highlighted in yellow. Or here I can construct QGIS to build the data according to the density of a population using HeapsMap. It was also possible to go back to the database and extract more information about the addresses and build an understanding of relationships with where the artist lived. So, here are all the artists for whom it was clear they lived on Charlotte Street for Trobia, which I'll explore again more in a moment. Now to the realities of executing such research. I've thoroughly enjoyed collating data and exploring research threads that I've never imagined would emerge, however, I face a lot of challenges. I face objections from within my own department as to whether the use of mass data was art history, but rather sociology. As Martin, my role has argued in making modern artists. Martin was formerly my supervisor, by the way. Art history does not exist without sociology and such a social context can be facilitated by large scale datasets. When I started, five years ago, I found there was a disparate information on how to collate the data. But since then, huge technological and scholarly leaps have been made with the likes of layers of London being open access and the publication of the Routledge companion to digital humanities and art history. There is the challenge of actually interpreting the data, which is why it's important to have a fully functioning platform. For example, I had to pay around for a lot of time with Microsoft access as to how to filter things by time period and the technical and historical knowledge to interpret that data. Initially, I struggled to find training and technical support, but I did manage it in the end and I needed training to use the likes of Python, OpenRefine, Microsoft Access and GIS platforms. I was probably too ambitious to be honest. A big challenge was executing the research with no project funding, but I adapted my approach so that I only used free open source software. However, in their present format, neither the database or map are publicly accessible. Because mass data hasn't been embraced in art history at all recently, projects are often on different platforms making comparison more challenging. Or they have different criteria for creating data, and thus aren't strictly comparable. So it would really be interesting in seeing how we can work together to draw our projects to make comparisons. Whilst these may seem like negative points really, I'm highlighting them because I'm so enthusiastic to iron these things out for other people. So mass data was only one third of my methodology, providing a macro perspective of the landscape artist community. The database could not account for the context of artwork's creation or production or its cultural significance. To redress this and balance the second stage of my research that zooms in on a meso level, that is looking at a cross-section of the landscape artist community through four case studies, John Constable being one of them. The third stage of my research zooms into the micro level of the studio, using qualitative material to understand the activity that's put within and moulded the landscape artist studio. Adopting a multi-scaler mix methodology was beneficial for three reasons. Firstly, the cross-sographical approach, using the scarce amount of data cumulatively. The second part of the significance of location was retained at every level of analysis. To keep the locations in relation to landscape production at the forefront of my analysis. And finally, the mix methodology allowed me to simultaneously ask questions of the overall landscape artist community and the individual experience. The macro data would just be statistics if it wasn't contextualised by the qualitative data. Likewise, the macro data provided an overview of the landscape artist community. The meso and micro analysis would have mirrored previous monographic scholarship, which provides an limited context for the wider production of their works. The findings of my research devise institutional narrative that there are only a handful of successful landscape artists and has revealed different communities within the landscape artist population. Particularly female landscape artists who scout and feature in literature. In short, this methodology provides an enhanced macro view that demonstrates the possibilities of alternative narratives that can be drawn from our state and GIS resources. So I'm now going to briefly present a case study to demonstrate the value in the multi-scaler methodology. Charlotte Street Fitzrovia has always been noted as a popular street for artists, however there's never been a macro analysis of who passed through the street or the significance of the genre. I've mapped 23 landscape artists who resided on Charlotte Street. Charlotte Street offered cheap rooms, which were ideal if the landscape artists were visiting for the exhibition season. However, at the other end of the scale, artists who were well provided by independent means such as Joseph Browning would build extensions to accommodate their practice. Browning lived at 35 Charlotte Street for 39 years, which was highly unusual for landscape artists. So the data collection was really important for reflecting the reality of how landscape artists typically occupied numerous lodgings in London throughout their career. Situated almost opposite Baratyn's property was number 63. The first landscape artists we are aware of lodging here is George Moreland. Moreland's notorious lifestyle of drinking and debt was prevailed in art history, art historical memory, more so than art. By 1792, Moreland was legally bound to pay his creditors £120 per month to consolidate debts and was housed at 63 Charlotte Street. It was described as an elegant house with a garden coach house of stables. When Moreland occupied the house, there were four of the landscape artists on the street, including Francis Wheatley, with whom Moreland had published numerous prints throughout his career. In the same years movie he did to Charlotte Street, Moreland exhibited an astronomical 15 works at the Royal Academy. He reduced his canvas sizes to increase productivity, but he's made some exceptions to attract attention. The benevolent sportsman pictured here was one of the largest works ever painted at Charlotte Street and was reportedly executed in about a week to ensure it was ready for execution. Bit of an exaggeration I can imagine. 17 years later, in 1811, John Constable would move into number 63. Whereas Moreland had had the whole property, Constable leased rooms on the first floor above the cabinet, make it an post for a rich weight. Constable requested that the back drawing room would be painted salmon with crimson post of furniture. Constable himself was a different batch show that all would assist upon it, reassuring himself that the colours will suit the pictures that he wanted to display. Constable's studio was the first floor front room, likely the space used by Moreland, which had the slightest barter from northeasterly light. Here he struck to the studio to be painted with a sort of purple brown from the floor to ceiling, not sparing even the doors or door posts for white is disagreeable to a painter's eyes near pictures. Arguing Constable's wall colour was strategic to create a neutral environment to not influence the palette of his works. However, in February 1814, he was struggling to add warmth to a cloud scene in Suffolk. The studio would not have had the ideal lighting conditions due to the combination of purple brown walls and the London winter. Unable to resolve the issues in the compiles of the studio, Constable pledged to prepare an on-plane air oil sketch from then on. There are no surviving images to see if it would be Charlotte Street or its interior, but fortuitously, the landscape of artists of figure painter William James Moeller drew the inside of his studio of 22 Charlotte Street. On the east side of the street, Moeller's first floor room studio would have had less light than Moreland Constable. In this confidence sketch here, Moeller places emphasis on the context of the room. Although it was likely that Moeller had several rooms to live in, this sketch shows the limited studio space an artist of Charlotte Street realistically had. Though this is a very brief and focused example, the database and map reveal relationships and circumstances that would not otherwise be possible without mass data. However, the data was only contextualised as part of the multi-scaler mix methodology that allows us to understand how a location could shape a landscape painting. There are far more questions that could be asked of this data set. If we consider the number of landscape artists identified in the database, around 2700 exhibiting over a 70-year period, then it's possible to imagine the burgeoning activity that was occurring within our studios and how many more are waiting to be uncovered. I designed this research to have a life well beyond my thesis. My methodology was designed to ensure internal validity. That is the process that can be replicated to generate the same results, as well as external validity, so the methodology can be applied to artists' studios of other genres to allow for direct comparison. I would like the database and map to become publicly accessible and user-friendly. The data itself will also be contributing towards collaborative research such as the British Art Network's Mapping British Women Artist project. So, looking to the future, mass data could be utilised to look across time periods, artistic genres and material practices to name a few. But I would argue that embracing multi-scaler mixed methodologies and advanced technologies are essential to art history if we were to expose forgotten artistic communities from little surviving material. I welcome anyone to get in touch to discuss my methodologies and resources I've used. Thank you. Thank you so much, Rhian. That was so interesting, and I think it speaks really brilliantly to some of the questions we had in the first panel about the kind of nuts and bolts of how people are working and connecting it through to specific examples of mixed forms of analysis. So, yeah, thank you. And it will be great to discuss your paper more in the Q&A. I'm going to introduce our second paper now, which is being given by a team of researchers who have very cleverly coordinated Zoom backgrounds. I'll introduce Ming Tiampo, who is Professor of Art History and Co-Director of the Centre for Transnational Cultural Analysis at Carleton University. Penzi Atta, an Egyptian-Canadian visual artist, curator and researcher living and working on the unseeded territory of the Algonquin Anishinaabe nation in Ottawa. Yanake Van Hoove is pursuing her MA in Art and Architectural History with a specialisation in Digital Humanities at Carleton University and Maribel Hidalgo Urbanecha, who is a Worlding Public Cultures Postdoctoral Research Fellow at the Chelsea College of Arts at the University of the Arts London. And their paper is Mobile Subjects Contrapuntal Modernisms. Thank you so much, Bailey. Are you able to hear me and see me? Hello. Yes, we can. First, I just wanted to thank the Paul Non Centre, Sarah Turner, Martin Myrone, Danny Convy and Bailey Card for pulling this all together. It's really such a pleasure to be in this context. Reflecting the importance of collaboration to data-driven work and also to global art history from a pluriversal perspective, which I call Worlding, we are presenting in a team of four today, although we're just four of a much larger team, which represents the digital side of these much larger collaborations. So I'll be speaking first, followed by Panzi Atta, then Yanake Van Hoove and finally Maribel Hidalgo Urbanecha. Mobile Subjects Contrapuntal Modernisms investigates the circulation of artists from the decolonising world through the colonial and artistic capitals of London and Paris. This tale of two cities considers how these capitals of decolonising empires functioned as critical meeting places, anti-colonial hubs and sites of exchange in the decades after World War II due to post-war mass migration. It considers how these mobile subjects, artists from Asia, Africa and Latin America, relocated to the different immigration contexts of London and Paris facilitated by scholarships, educational opportunities, exhibitions and journals. It furthermore seeks to make visible the transcultural artistic practices, communities, networks, discourses, solidarities and worlds that artists co-constituted through, between and in these capitals, as well as their impacts on regional histories of modern art. The purpose of this project is to propose a new analytical model that sees metropoles not as points of origin or as global training grounds for artists from the global south, but as spaces of intersection and flow that allow us to understand the transnational condition of modern art and the co-constitution of modernism. In order to do this, it is not enough to rely upon the traditional methods of art history, nor on any one individuals into the expertise and data sets. So here, as Paris Spice Gans argues about women artists in the 18th and 19th centuries, it is necessary to demonstrate that these artists from the global south in London and Paris were quote, exceptional, but not exceptions. The project thus uses data science to connect scholars and their existing data sets through a collaborative database and workshops that map and investigate the presence of overseas artists in London from 1945 to 1989 and Paris from 1900 to 1989. Working collaboratively, and our team is much bigger than just us, we are experimenting with artistic visualizations, engaging with issues of decolonizing data ontologies and database structures, and are now working to bring it all together in a critical database and data visualization, which we will demonstrate later. As Diana Sieve-Greenwald argues in her book, Painting by Numbers, data science is only one tool in our analytical toolbox, which our team uses as a starting point for art historical and interdisciplinary methods that range from formal analysis and slow looking to postcolonial and decolonial theory, global micro history, global urban history and worlding. As one part of our larger methodology, data provides a macro view, something that we've heard many times today, allowing us to zoom out to see patterns and connections before zooming in again to look closely with fresh eyes able to see beyond the disciplinary assumptions of art historical inquiry. And this is very similar to what Rhian just presented earlier on multi scalar mixed methodology, so I think I think we have lots to talk about there. So what I'm showing you here in the slide is an example of what our colleague and team member Michelle Greete did with her website and book Transatlantic Encounters Latin American Artists in Paris between the wars and so she has a book, which is much more traditional piece of art history, and also this website which I'm showing you on the right hand side. I'm here is a similar pairing of more traditional scholarship and digital methods in a collaborative article and archive feature Slade London Asia that I created with Liz Bruchet in British Art Studies. Now I'll put the links in the chat later. I'll be using a large data set as a starting point. The article in British Art Studies I'll address large movements and trends and also zoomed in to explore micro histories, such as what the ways in which artists from decolonizing contexts went on to establish and consolidate art institutions in the global south in critical dialogue with their time at the slate. I think we have all of the smaller micro stories the micro histories that we presented in the context of this larger macro history in British Art Studies. So now I'm just going to hand it over to my colleague pansy. Over to you. So when we when we that we've sought to visualize these concepts is through this animation, tracing the movement of international artists to the Slade School of Fine Arts through the mid 20th century. Unlike the standard Mercator projection, which is typically used to visualize the globe. This map is centered on the Pacific rather than on Greenwich. This draws attention to the global spam of these artists networks shifting London from the center to the margins while keeping the relationship between the continents familiar enough to remain legible. At the same time, geopolitical borders are rendered as organic porous indeterminate charcoal lines, making visible a semblance of their social and political presence, as well as their volatility over the course of this time period. Similarly, the paths denoting the artists movements are sketchy indefinite multiple their lines produce a geographic and temporal network that is transiently interlinked through sharp digital artifacts that branch out across the maps clusters. Animated using blenders open open source robust tools, each of the paths visual characteristics, their opacity texture their visual noise can be generatively mapped on to the data sets variables encoding seemingly decorative visual qualities with quantitative or qualitative meaning. Accordingly, this work reflects and reproduces the project analytical methodology. While we know the artists national origins and their destinations, we cannot be certain of their routes for the multitude of possible connections they may have made along the way. The materiality of this visualization puts into practice a decolonizing approach to mass data methodology by making space for multiplicity and uncertainty in analysis of the global, telling a story about the numbers that accounts for the indeterminacy of the human. And now over to you know this work. Thanks pansy. Okay, so what we're looking at here at its core this network model shows approximate connections that existed between racialized artists beyond this the model can help us identify new insights and evidence. It can help us identify blind spots as described by Johanna Dricker in our data, particularly in areas that have been historically overlooked. My view does not tell us much other than reaffirming the fact that there was a massive network of non Western artists active between Paris and London throughout the 1900s. The colors are coded by country of birth and there is a dot or node representing each artist in a second node connected by a line for each connection point or institution exhibition or gallery that an artist was involved with. There are 3073 artists in 2993 connecting lines within this network. Next slide please make. So when hovering over an artist a box will pop up with more detailed information. We are mindful of the gaps in our data, as indicated by the question marks beside the fields that we do not have information for. From here we can also identify some major clusters within our data, including was it a international and slate school. Next slide please. And we're playing a video here without sound. So to develop a better understanding of the stories buried within this network, we need to dig deeper into the layers of further stuff we have taken us to prepare a second version of this network with a time based animation to visualize how this web of internet has fluctuated and evolved. As you can see the network builds over time with new connections each year being represented by the colored lines coded by country. We could also isolate a specific timeframe or country of birth using filters, forcing individuals connections changing over time within this network and I'll give it a minute to finish playing here I talked a little bit too fast for this video. Connections are still building over time and eventually we're going to end up with view that shows us at the end of the 1900s or the midpoint. The amount of connections within our network. Okay. Next slide please me. Another filter we can use to dig deeper into this data is filtering by specific connection point. So this static network includes the same information about artists who studied at the slate school that pansy just presented in her animation. We can see all the artists who had an affiliation with this late school over a span of roughly 80 years, which still does not tell us much about the possible connections between artists. Our next steps will be to create more flexible code based models of the data that resonate with the uncertain and inconsistent nature of the histories we are working with this data contains gaps, ambiguities and multiple meanings that situates these histories in a fluid rather than final position over to Maribel. Thank you, Jenica adopting a quantitative methodology and in our case, a response to the need of a question in stereotypes and assumptions, which is a principle that guides data centric digital humanities methods. And these approach answers questions that require evidence of a different scale of complexity than traditional our history challenges monographic studies. But in addition to this key to our project is a critical intervention of our methods. The dialogue with the decolonizing data research cluster of the word in public cultures project mobile subjects contra puntal modernisms is engaged in a critical interrogation of data centric methods. If you're curious about the word in public cultures projects and and also about the work we've been doing and around the seed oxyram ontology. You can visit the two links on the slides. Critical nature of our approach is leading us to up to adopt post colonial and decolonial digital humanities strategies and practices that respond to the fact that, as Rupika reason argues, cultural canons are being reproduced and amplify not only in the visibility and discoverability of knowledge, but also in a post demologies of digital knowledge production. The adoption of well established and universally recognized methods that are often seen as neutral by the white portion of the community is considered good practice as it facilitates a change of information but the truth is that the community has reinforced 16 power dynamics and knowledge biases. So in response to these issues and quoting Rupika's reason work again post colonial approaches to digital humanities intervene gaps and omissions by engaging with the politics of representation and look beyond representation to develop design practices that lay bare the politics surrounding digital knowledge production. Accordingly, our aim is to address gaps and omissions in terms of artist representation, as well as to rethink digital knowledge production and epistemologies. As we develop the project database and data visualizations that accommodate artists as next next slide please. In particular and concerning data visualization, we found ourselves engaged responding to the idea that modeling data into a graphical expression such as a visualization is a form of knowledge production and a form of its design, namely an in connection to the visualizations my my colleagues pancy and Janaka have just shown. Johanna Drucker states that maps are a record of exploration pathic and tactile narrative and immersive when creating in from inside experience of discovery, but rationalized through projection when produced from outside as images. I'm representing a curve form on a flat surface, as well as a many cultural imperialism work, provide their own history within the range of projection methods. And then about network diagrams. I assume an absolute distinction between notes and the edges that connected and their displays organized by algorithms that optimize display rather than following strict interpretation of their relationships between among the nodes. The visual arrangement cannot be read as an accurate presentation of information only as an approximation. Next slide please. I would like to finish our presentation with a reflection on the next steps that follow critic and acknowledgement of the piece demologies and knowledge production. And we'd like to share with all of you the following questions, which are guiding our process of rethinking and redesigning the database and data visualizations model we use and and these questions are. Can we decolonize data centric methodologies. How can data methodologies be reimagined to render complex histories and identities and to surface horizontal relationships. How can we challenge dominant models for data visualization visualization that have been built upon colonial and western epistemologies. I would like to express uncertainty and be with the contradiction and cultural nuances through data visualizations we will really like to have to know your thoughts about these questions during the Q&A on to finalize. We wanted to say thank you to our amazing colleagues without whom this project could have never been developed so thank you very much. Thank you so much for that fantastic paper I see lots of applause. And for those questions which we will come back to and it's really exciting to see all the connections between our first two papers already as well. I'm going to introduce our final speaker, Mary Okin and congratulations are in order because Mary's PhD was conferred today. I'm very pleased to introduce Mary as holding a PhD in the history of art and architecture from UC Santa Barbara. Mary is also a lecturer at San Jose State University specializing in the history of American art and digital humanities. Her paper will be challenging the American art canon with mass data mining at 10th Street visualizing New York City's 10th Street Studio building. Thank you. So good morning everyone. I really delighted to be here. Can you hear me? That's great thanks Mary. So I'm really delighted to be here and my presentation concerns a digital humanities project that I've been developing as a side project my dissertation and it's still and it's kind of nascent stages. Although we've just myself and by former student and unpaid research assistant have just published an article recently and I'll share in the chat with sort of preliminary findings. So I suppose I should say that we're still in search of a method but we've explored some and I hope that'll be interesting for people. So I am a California based Americanist with a specialty in histories of California painting. And my research on California painters has made me curious over time about why it's the case that the measure it seems for determining an artist significance in place in American art history is often bound to whether or not they achieve fame in New York. It seems to be a kind of pattern that you have to have an exhibition, whether you're living or dead artist to be considered important, and that New York is that kind of central place. So this project, which developed out of a graduate seminar where I was kind of, you know, you need to find a data set and explore it has led me to really think about that and come up with at least some answers for the origins of that center of art being New York. The project has also helped me to challenge or explore the single author model so it's lovely to see that among colleagues as well, and to think about, you know, how to be part of a community that is experimenting with data. I've had the great fortune to be participating in the early stages of this project in something called the Tour de Mour Inclusive Digital Art History Initiative that was started by Panorama the Journal of the Historians of American Art. So I've had a chance to work with Diana C. Breedwald and other folks and it's been amazing. So thank you, both to them and to my unpaid wonderful research assistant, Seely Richard. So the 10th Street Studio Building for those of you who may not be familiar with it was the first building specifically designed to house artist studios in the United States. It was developed and built in 1857, and it was demolished in 1956. This visualization in the middle is a mapping visualization showing some of the historic structures, including the studio building that are related to its patrons. And so what you see here on the left is the studio building at the very top, which I'll show you a larger image of in just a minute. A stable and then the house of the first owner of the studio building, a man named James Bormand Johnston, this is his house on the left. On the right is the mansion of his older brother John Taylor Johnston, who was a very important American art collector, and also the first president and one of the founders of the Metropolitan Museum of Art, which was founded and meeting at his home around the quarter and a couple of blocks over from the 10th Street Studio Building, which was founded or begun about 13 years earlier so 1857 for the studio building and 1870 for the bat. One of the things that we've been working on is to define or maybe redefine the studio building so the most common definition, I would say in the art historical record and among the art historical community is that it's a home for artists. What you see in the image on the left, the building of the first building specifically designed to house artists has large windows, three stories of studio space that were ideal in their moment of creation. They were modern in a sense of artificial lighting, really large scale space with northern facing lights, a large front door through which, you know, grand works could circulate, and a two story gallery in the center where works of art could be exhibited. And for the first 50 years of its life or thereabouts, it was kind of a social club and creative space primarily housing important male artists and the image on the right, a stereoscope photograph is an example of that kind of community. So this is a photograph of Worthington what churches studio and some of the prominent male artists, many of them landscape painters who were living in this building and working in close proximity to each other. In talking to Diana Greenwald about this project she also works on American art during this era especially. She suggested thinking about this building in economic terms as a creative cluster. So thinking about the fact that artists are interacting with each other that this building, which I had already come to understand is an economic engine of a kind for developing the American art market and the American art community and a period where that's really a fledgling enterprise in the United States. And so we began to kind of think about that further. Some of the artists, the major artists really early on in this home for artists are people like Frederick Frederick Edwin Church, excuse me and his painting Heart of the Andes on the left, which was the first blockbuster exhibition of American art which took place about a year after the studio building opened for business. And on the right is a painting by William Merritt Chase, an American impressionist painter who took over the large two story gallery where Heart of the Andes had this blockbuster showing and turned it into the largest studio in this building, which was the kind of social hub and so on. If you walk into a room of 19th century American arts, this is a room at the Metropolitan Museums American Wing, you are going to encounter paintings by folks who lived at the 10th Street Studio building. And so it's very central to the canon of American art, and it's interesting to study for that reason, not only to explore. It's very well known artists but as I've been doing and bringing in more peripheral figures and trying to think through what they mean in relationship to what we already know about this structure and social network. So here's a Google and Graham visualization of the 10th Street Studio building as a proper noun. So to look at its presence or frequency in English language publications and where it shows up relative to the two artists I just showed you works by Sir William Merritt Chase and Frederick Edwin Church, as well as its two first owners so James Borman Johnston and his older brother John Taylor Johnston. You can see that John Taylor Johnston has this kind of peak of recognizability during the period that he's the president of the Metropolitan Museum of Art, which you see in an early photograph at the top here. If we look at a Google and Graham visualization of the 10th Street Studio building so zooming into that from the visualization I just showed you, you can see that it starts to appear in the art historical record when it's first established so in primary sources from the 1860s. And then during its sale and demolition period in the 1950s, and then as it becomes of interest to scholars particularly in the 1980s. And in the 1990s when Annette Blaugrand, who is the leading authority on this structure began to develop her dissertation and the exhibition catalog that came from that work. So she is the first person to think of the 10th Street Studio building as a data set, and to create a list or roster of tenants that she published and began to publish even before she completing her dissertation so in the early and mid 1980s. And she assembled a data set that listed all of the artists who she could identify who lived in the building, and that began to create an understanding of this community at a macro scale over the course of the first half of its life. This project, the mining at 10th Street project began with that kind of foundational data set so that class I was taking where we needed to find a data set begin experimenting with it. Her data set is pretty straightforward and limited in the sense of what she published. So it includes the first and last name of the artists their life dates their years of tenancy, and then their affiliation with the National Academy of Design, which was an organization with close affiliation to the studio building its patrons and many of its tenants. And I began by asking, you know, what could we add in terms of data, and how could we mine additional data that might give us further insights about this canonical community and building. So over the course of the last four and a half years, primarily using directories professional directories and city directories we've added substantially to Annette's data so her data is represented in green. She identified 159 studio building tenants, a majority of them fairly well known artists, and we've added another 260, probably more than that at this point, but in this snapshot of last year about 260 artists. A majority of them falling in the latter portion of the building's life, and I'll say something about the three eras of ownership. And the mass data project began very simply, which is just a kind of data validation process, which is to say, you know, trying to find where did she have where did she find her data and could I living in California far away from New York archives. To validate her work and sort of like a data scientist you know test what she came up with us her conclusions and so on and see what else I could find. Of course it's been since the 1980s that she conducted her research, there's a lot of digitized sources, you know it was exciting to see what what I could find. So very simply I Google searched for the address of the studio building and immediately city directories popped up and I began kind of searching for the building's address. And that led to also looking through periodicals and very quickly a much more complicated picture of the studio building emerged. There were some challenges with that immediately it was clear that freely accessible sources were very limited and many of them unavailable, and that led me to you know contacting the major resource for directory data for New York City, which is the New York public library. That institution has in their collection city directories for the city of New York, and they've digitized them and they're available to the public. However they're not available in searchable form. And so I contacted them to ask because you know going page by page and searching in a kind of analog way for a studio building address would be a gargantuan task that I don't think anyone has time for. So I contacted them and found out that they licensed their directories through ancestry.com and it's subsidiary fold three which is a database with military records and city directories and other kinds of works. Not all databases that's one of them are particularly user friendly so we had to come up with techniques for how to locate, you know how to search genealogical databases, for example for a building rather than a specific individual which poses, depending on the way that database works, a lot of challenges. And then we had initial findings so we discovered more artists and architects, many of them obscure non artists tenants, so adding professional diversity to the studio building which city directories especially provide, because they are less discerning than art world records for example. Women artists and women in general began to appear in much greater numbers particularly when we began to extend the data set into the 20th century, and that's work sort of stopped at 1895 or so. And then we found organizations and businesses that led us to think about connections to capitalism and the development of New York City, not only as an art center but as a center of finances or the financial world for New York City. And the role that the patrons of the studio building who were major business owners and business men with power in terms of economics and politics and the studio building as a phenomenon that's cultural and primarily thought of as this home for artists. We also discovered domestic staff. I think that's really interesting and important in thinking about the art world and the community of cultural production in general that there are people who, you know, support that work in case the studio building Margaret winter who provided lunches for example for the artists. Data assembly I'm sure many of you who work on this this isn't any sort of surprise, but we searched base for basic addresses and then had to come up with lots of variables for how they appear so part of the project is documenting all the ways that you can find the studio building in the data. And similarly for the proper noun version of it which varies as well from newspaper to newspaper and among different writers writing about it so 10 Street Studio Building the old studio building 10 Street Studios and so on. Some of the exciting findings for us were the organizations and businesses especially. So the studio building association was something that came up in the city directory and I don't think it's available or prominent in art world records. And that was an organization and a business that was founded by James Bormand Johnston, along with a number of other patrons, and including artists of the studio building who were members of the patron class. It's listed in the city directory is having capital worth $500,000, which I tried to translate into today's money in dollars and it's like 10 billion or something. I don't know exactly what the translation is but it's a lot of money for the 1860s. And that was really interesting in finding the corporate document related to that, seeing how the patrons of the studio building were really supporting the art market and doing all they can to promote the artists of the studio building but also New York City as a center for American art and hopefully creating a kind of Paris or Rome with this particular project in New York City. So coming back to that idea of the patrons. I hope I've complicated that a little bit that there's this money in connection to their appearance and presence in the historical record. And that becomes apparent to when we add the Metropolitan Museum, for example, to this, this visualization. So John Taylor Johnston that like high peak in this visualization is this much smaller mountain in comparison to the Metropolitan Museum of Art, which this family also built and that was has outpaced the studio building and that part is really important, because the studio building wasn't just a cultural center, it was a business, and it was treated as such whereas the Metropolitan Museum became a cultural institution the studio building business. In fact, it was a real estate asset and a real estate venture early on that the liberal proprietor as one of the newspaper says, without any anticipation of profit was now receiving 10% interest on the money invested in this is within a decade of founding a studio building. As time went on, after the first two owners so James Bormand Johnson and his brother John Taylor Johnson who were big patrons of contemporary art. As they, you know, left the world and left the studio building to John Taylor Johnson son, who had less interest in contemporary art, the studio building sort of reverted more to a real estate asset than a contemporary art center. Many of the famous artists began to leave. And as a result, there is a kind of shrinkage of its appearance in the art historical record art art critics were less interested in its tenants and there was a less what inviting or appealing reason to be in the building if you wanted to be at the center of American art. And eventually, the building was sold to its artist tenants, a cooperative was formed, and that cooperative eventually dissolved because shares were sold and eventually was sold to real estate developer and demolished. One of the pieces of advice that Diana Greenwald gave us was to consider in terms of telling the story of patronage of the building as a business of this like economic history of the art world and its relationship, meaning the studio buildings relationship to the American art was to track accessions of works of art by tenants of the studio building, and what you see in this visualization on the left are the three eras of ownership, and the number of accessions of works of art by tenants of the studio building split along gender lines so at the top you see the era, the first era of ownership of JB and JT Johnston and the number of accessions of works by men and women of the studio building. And the decrease of that over time, as well as the increase of the presence of women in the 20th century which is something that our data set has documented in the sense that more and more women come to occupy the space. It's time that the studio building becomes of less and less interest to art historians, art critics and so on. So, we track sort of this, you know the formation of the building as business, it's fall from grace in the, in the art world, and some of that maps against data losses as well. The data rich era of the studio building is that early period of time when it's connection to American capitalism and patronage is really strong. And increasingly there's less and less data available. And so the diversity of people that we've been able to find sort of shrinks a little bit. And that's one of the challenges that we're continuing to pursue is the 20th century data that is going to continue to hopefully emerge from more digitized directories and whatever else I can find on my next trip to New York which I hope will, will gather more data, more access to data as well. So, as press coverage dies down, there's less and less information from the art world, the retirement and death of the artists in the second era of ownership who were the most famous also kind of leads to data losses and then finally, the artists cooperative has the least amount of sort of presence in in art world data. And that reflects of course, what happens when patrons completely disconnect, I think from a contemporary art center that now is no longer the center but a periphery of American art. It's interesting for us because New York City and Greenwich Village emerged as the center of the art world, internationally after the war, at least from the American perspective. And so the fact that the studio building is right at the heart of that, and yet also cultural periphery is allows for the exploration of tensions between center and periphery in a very close proximity geographically to, you know the things that we think about abstract for example, and its relationship to this 19th century structure. I wanted to show very quickly, I hope I'm okay on time, very quickly, some of the raw data and future directions for the project. So this is a visualization of data that it hasn't been cleaned about professions that city directories register. You can see the kind of diversity or range of types of individuals who were tenants. This is a little bit deceptive because a single individual can register multiple professions and so some of this variety is, you know, connected to one person or a few people as opposed to many. But this shows you some of the kind of general categories in which tenants fall so artists being one of the largest architects and sculptors not that surprising, but also merchants and lawyers and illustrators and businesses that give a much more complex and interesting view of the studio building and what we can think about and and explore within its history. We've also continued to pursue the story of its patrons and their social network which is going to be one of the areas of expansion for the project. The father of John Taylor Johnson and James Bormand Johnson was a prominent Scottish merchant, who, among other things imported linen from Scotland, which I think has a connection to providing materials to plantations in the south where cheap linen was used to clothe the enslaved, as well as railroad and steel and so larger histories of industries in the United States. I also have started visualizing some of the marginalized communities so this is a story map and I'll post the link in the chat rather than showing it to you of the women artists of the studio building who we've discovered and the kind of international and professional diversity and connections that they have to histories at the studio building. I will post in the chat, some of the publications that we've started to create in case you're interested. Thank you so much. Thank you, Mary. Thank you so much. I love that at the beginning of your talk, you shared that you came to this subject through an interest in landscape painters in California. I think those kind of research trajectories are really fascinating and it, it's really helpful to kind of know that that connection. It's a very marginalized genre in the United States. Yeah, exactly. Yeah, exactly. And I think, I think we should pick up on on that as a theme from these papers. Also encourage anyone listening to feed us questions for the Q&A space that we have now for about half an hour, at any point, if you would like to ask the panelists something. So coming back to this sort of very clear through line through all of your papers around canonicity and challenging those very rigid sort of inherited boundaries something I'm really curious about is how you think about the scope of your projects I think that came up in different ways. Obviously something about working canonically is that you inherit a sort of container for your work, whether it's the life of an artist or people who are considered to be part of a particular movement. So on to that in your research, how have you each dealt with questions of scope in a very practical sense maybe for your work and, and, you know, maybe in a kind of evolving way. And so I've already, you might have seen I've already briefly answered this in my chat in the chat and the, I should say the most captive PhD, which Martin and my other supervisor Richard Johns proposed, and I applied for it, but I got to define the scope. So it was, I was looking, I decided to focus on landscape art studios. I had to define London, which was not fun, so I used a map to really set myself a limit and 1780 to 1850, 1780 was the year that Royal Academy moved to Somerset House, and it really picked off the first, the age of exhibitions. 1780 to 1851 was the great exhibition at Crystal Palace in Hyde Park. So that, for me, marked the beginning of the second age of exhibitions, so 1780 to 1850. And I have, yeah, like I said earlier, huge appendices, like detailing how I defined a landscape painting, a landscape artist, like all stratified samples I used, because if I wasn't strict, I mean my data set was huge as it was. But I had far greater ambitions, and I've probably done about a tenth of what I originally proposed. So, yeah, you can keep going forever, but I was really interested in what I did spend a lot of my first few chapters on my thesis just to find the use of what is a studio, let alone what is a landscape art studio, especially in an era much earlier than Mary's research where the records are very thin on the ground, people mainly worked in their accommodation. How do we, how do I count for that? So I just say landscape artists addresses, you know, they aren't all necessarily studios, but that's the closest we're going to get unless more information comes to life. Thank you. Ming. Thanks, Bailey. So, our project is actually incredibly fluid and enormous and quite difficult to get a handle on, but there are different parts of it that have more robust definitions and boundaries, one of which is the Slade London Asia portion of it, which, as you know, actually emerged out of a fellowship that I got at the Palm Ellen Centre, and like Rhian, I applied for this fellowship to work on London Asia within the context of pedagogies, and so that's when I started working on Slade and, you know, found this incredible data set, which, you know, had to be extracted very, very laboriously, you know, just by looking at each of the registrar's records and recording names. But then, you know, I think that in thinking about the question of scope, it's really useful to have the multi-scaler approach that Rhian was talking about earlier, where, you know, there are different levels of engagement with the data in the project, and I think that, you know, some aspects of the project are really just to provide this larger context. And then other aspects of the project allow us to tell sort of more precise stories through a kind of, I mean, for me, what I think is really important is this concept of global micro-history, that you find a very small site that allows us to tell stories about transnational global engagements from a critical perspective. So you have that kind of precision, while also engaging with larger questions about global. So I guess that's sort of an answer that sort of tries to have the best of both worlds. But I think that, you know, this question of the multi-scaler is really important. Thank you. I think something that came up in those answers as well has to do with the idea of sort of data ontologies and how we're kind of managing data sets and what is reflected in data and what isn't. The question of definitions for terms, what you incorporate into your set and what you don't. I was wondering if you could share maybe some of your kind of tactics from your projects for dealing with what might be some of the inherited biases and how we might normally structure our historical data. I think for me, I think for me, including, you know, folks like domestic staff and Lord and trying to sort out what are they doing in this building. You know, merchants, for example, or an undertaker. I think, and then figuring out how to code, which we're still trying to sort out these various professions and how do you quantify and explore the data that way. I think also, you know, in some cases, names change slightly so also how you identify individuals and working with the large data set you can have all kinds of errors that pop in. Which is part of the fun, I think of sorting that out. But it makes, I think there's all kinds of decisions that that you can make and how to tell the story and I think I'm still trying to sort out and so it's just mentioned, you know, where the limits should be because the data set can expand in so many directions. We're thinking about education. I didn't show you a slide, but home addresses, for example, pop up in the directories and home addresses in New York City between 1857 and 1956 and the maps we have today are like wildly different. So, you know, trying to figure out how you geolocate them, which I think would be really fascinating. I hope that answers that question a little bit. It does and it also picks up on an interesting thread through the day, I think, which has been the value of errors and anomalies and how things that stand out or that don't fit in your data can actually point to things that really need attention. Paul, I saw you raised your hand. No, it's actually related. Sorry, but I think Rheon and Ming want to respond. So I'll wait for them. I'll defer to them. Okay, I'll come back to you. Rheon. And yeah, my biggest challenge was, and this partly the motivation I understand for the Rhe Chronicles 250, was that so much of the exhibition Street of London is based on the out on graves dictionaries. And they were my go to resource. I quickly realized that I could not go to all the exhibition catalogs individually. And this, by the way, I was doing my take a collection six months ahead of the PMC digitising theirs. So if I've done my PhD two years later, it would be so much easier. I have to do that. I like a lot of go through all the grace dictionaries deciding is that Alaska part is by my definition or not. And, but graves has his own errors and biases. And so I really wanted to make sure from institutional perspective that I was cross referencing from other institutions. And that's why I was so heavy on tracking all of my processes so my database includes it down to the page number. It finds a contradiction in the future brilliant flag it because I'm sure my own, you know manual process has meant that actually I've probably made errors in the data in applying my own data. And but what I really was concerned about but I've had to just leave it as kind of a foot move aside project in the future is that as much as I've raised the issue of gender throughout my database. It doesn't address race and class and the colonial structures which actually are already within the initial cataloging of the artworks, and the titles they use, and the purposes of their works, the motivations of sending the text. So they're proposed another 10 PhD is just of all the different directions. I get the sense that is a part of work like this that you have to, you know you're identifying horizons around you all the time and there's a kind of sense of selection as you go maybe as well. I've had to accept that I just can't answer everything. So, if you create sending cut projects that other people can answer an invitation. Well I'd like to just use this opportunity to maybe open up a couple of things and then point to other members of our team that might want to take on the answer. I think that for us, there are two places in which you know we're really thinking about how to address these questions of inherited biases. One of them is through the data itself and you know as you just pointed out, Mary, there were just so many places in which we were making assumptions and, you know, also what you were saying earlier around, just in terms of like, how do we identify who the foreign students are, you know for certain periods of time in the records, the slag kept very good records of who was coming from where, but otherwise, sometimes we really just had to rely on names that did not sound anglophone which meant that we were making giant assumptions about who was or was not a foreign student and we missed out on a lot of students who had anglophone names because of those colonial histories. So, you know, over time, we had to find other ways of tracking and finding those students, but then also in terms of the visualizations themselves so I think pansy probably has something to say about errors in her visualization at same with Yannicka, and then Mary Bell probably has something to say about this question of data ontologies. Yeah, thank you, Ming. I think that's definitely something I can speak to. I mean, for me it was less so about defining or like circumscribing the outside bounds of the data set and more about working with the data set to think differently about the sort of like to think kind of against the out of the box solutions that I think tend to be quite hard and sharp and non-horus and linear in a way that I think tends to produce like a flattening of the information that is being presented like via a visual model that implies a sense of certainty that I think is sometimes not necessarily there, you know, and that's a way that the stories that the data tells can get I think flattened out and lose some of the nuance that I think is so vital to taking this kind of decolonial approach. And so for me it was really about working with the data set we've had to try to re-inscribe what I think was a really core part of it which was this sense of you know that we are trying to find poetic, geopolitical, social, historical links that are, that have the, there is a potentiality of them being there that is pulled by the data that we don't necessarily know. And so all the different, like I was saying in the presentation, all the different variables associated with the mind's visual qualities can be encoded with quantitative or qualitative data. And that can be a way of, you know, like it's, this is I think what contributes to the sense of kind of materiality of the visualisation, this kind of like charcoal drawing look is because of this way of trying to acknowledge it and in fact put it at the forefront, the kind of the margin of error, so to speak, in the data set. But that requires really using like not out of the box solutions, I mean like, you know, like I was using Blender, which is incredibly robust, but it's quite hands on in some ways. And so to me that's like a pretty core part of like countering one of the biases of like visual presentations of data that can quite I think is always reductive. Am I allowed to jump off of what pansy was saying. Absolutely. Okay. Yeah, I also resonate with the thought of using out of the box. Visualisation platforms. A lot of people on this call were using Tableau to visualise their data. I did that too, with the network visualisations but I kind of had to use a few different platforms, like Jeffy to generate coordinates for the network visualisation before putting it into Tableau. So even combining different systems to get an end result that you want to have is there's some extra steps that need to be taken along the way and the major frustration that I had when I was making this network visualisation was making sure that we had the time element as a part of it because the networks like everything in one view doesn't tell you much if you don't have the time element to it to tell you which artists knew each other at which time. So we needed that extra level of information, but I had to be very reductive with how I used the times in the network visualisation. I mentioned some artists we had very specific like day month year for when they were at an exhibition other artists we just had a single year. And I had, or there was like a lot of uncertainty about dates for example in our history you see a lot of circa this year, or a range of years that an artwork was produced over. So to deal with these types of dates, it's difficult when you try to make a visualisation in a platform like Tableau because they just want a strict date. I don't know who I'm saying when I say day but the platform just wants a strict date to be put into the system, otherwise it just doesn't work you get a lot of null values and the data just doesn't look like it. It works in the end so what I had to do is I had to reduce all of the dates in our data set to just a year, which I felt really bad about like I was sitting behind my computer playing my hair being like oh should I really do this it like gets rid of all of the information that we that we have in our data but it was necessary to create the network visualisation and something that I've been doing in a recent project is I added an extra category called datedness to my data. And this was a set where I had a lot of data that had circa 1960 for example so what I did was I took apart the circa and the 1960 part I kept in 1960s and the strict years and all the data that I represented and then in the tooltip or the box that you can get when you hover over a data set. Over a data source, you could see the what I was calling the datedness of the data and let's see like it was circa 1960 or it was part of this range of dates so you have to try to be kind of unconventional of the way that you present your information. I love that idea of introducing a new dimension in the data to indicate degrees of certainty right or degrees of possibility and working with that instead Maribel would you like to jump into. Yes, and also I see that there is a related question on the Q&A about specificity basically the ontologies and how we are dealing with that. I mean, I will speak more about what we are doing in world in public cultures, which is an approach that we are going to adopt probably in these projects because we basically we decided to adopt the CDOC CRM model, which is like very, you know, well established recognise also we have a member in the project who belongs to the, to the, to the sake of the CDOC CRM. And I mean it's, it's quite problematic it doesn't, you know, it doesn't accommodate all, all types of data. It's, you know, it's a reductive is it basically represents a meaning it's a representation of a certain view about cultural heritage and and and arts and everything so we started with basically with trying to expose and critique the biases that are in the CDOC CRM and in the classes and properties that we are using. Our plan, we also like did this very collective in a very, very kind of collaboratively collaboratively collaborative way. So, so basically we have like sessions in which we were like looking at the different classes and properties and and with our, you know, from all different backgrounds because a word in public cultures project has many scholars coming from, you know, from many different parts of the globe and also specialising in many disciplines and areas so so we've been trying to highlight basically the gaps omissions and problems that exist in the CDOC CRM and our plan is at the moment is just like to highlight in our database once it gets developed like visually have warning signs pop ups that when you are like checking the database you can see that the problems behind the schema and I mean that if you if we really want to decolonise the CDOC CRM will you will probably shoot basically communicate what the problems are with the CDOC CRM and ask for for more like kind of, I mean like like a stronger intervention and changes and because I mean it's of course like we need to to use it because of because this is the way we can share data with other researchers or their institutions, but at the same time there are so many limitations and problems with it so the first step of course is like highlighting what the problems are on the biases, but we need to take a step forward and make like more drastic changes in the CDOC CRM. Thank you so much. That's so yeah that's really interesting to hear about and maybe that that before inventing an entirely new system is a process of like critical annotation maybe within the systems that are being used currently and that that is a level of research and analysis as well within the project. Paul, would you like to to jump in. Mary did you have something else you wanted to say or on that line. It's just going to quickly add I forgot to say that that one of the findings that that one of the things that the data doesn't show maybe from for my data set. And that's difficult to find is ethnicity, but I think generalising from what we have been able to see one of the interesting things about the studio building project isn't seems to be a kind of case study in whiteness and and therefore the issue of exclusion in the that's one area of expansion that I think will be interesting given how closely related the studio building is to, especially in the 19th century to the canon of American art. So I just wanted to respond that that's that's a very that's a challenging and I think also really worthwhile aspect of these kinds of data sets to person. Thank you. Thank you, Mary. Now we'll jump in if I may. Is that that's really it's really interesting discussion and it makes me think particularly of someone like Jessica Johnson wonderful article on markup bodies which is a kind of critique of the transatlantic slave trade critique and a complement of this very difficult and very interesting history. But I want to I want to pick up on on mings kind of challenge of decolonizing this and level of kind of raise a question about the resolution of our data in the sense that what we've all done and this is very much a critique of our work to is we've kept the resolution of the biographical. We've kept it at the level of the individual artist at the individual patron. I'm wondering if that's part of the problem right if you're working with the Benin bronzes you're never going to know the artist you're just not or the you know and whether it's one artist or several artists. If you're working on other kinds of periods. So so some points to you know the the biographical says the man working with the dictionary of art historians the biographical really limits us. In terms of decolonizing and so I guess it's really is an honest question of can we imagine a different resolution. Is it the group is the community are there other ways in which we can confront, which is clearly a real bias in the very nature of our history itself. Brian, would you like to answer that. Oh, it's good to just respond and say I suppose that then takes us on to what is art and how do we value it and then what culture culture perspective are we coming from. And Brian Westman on the entire value is on the artist well predominantly on the artist themselves isn't it. And then you get the myth of the artist and so yeah I don't have to go for that. Yeah. No go please go ahead. I'm trying to find something for us. There was a database that I heard a presentation about that actually was focused on the objects on a group of our objects and and allowing those objects to sort of animate communities. But let me just. I'm just going to pull myself out for a bit and I'm going to look for it. That sounds good. Mary. I was going to say you know for some of the 20th century data for the studio building a lot of the people we identify or the individuals we identify are obscure so while we have their names we don't know much about them their relationship to the art worlds, or the studio building. And I think through them, and just thinking about all of the other people who circulated within the space so students for example for somebody like William Merritt Chase. There is that kind of, I don't know haunting of the cannon and haunting of the what we know or don't know. You know the other thing about directories which is where kind of we've started and we haven't expanded like Annette did into artists diaries and letters and whatever you just get an annual snapshot. So it's, you know, it's such a superficial view of what actually is going on. And the mess of the art world on a day to day basis. So I think that's where data losses are really interesting of how much do you not know and is there a way to measure that and sort of evaluate what the data losses indicate about our assumptions as art historians of what's important and what isn't. And maybe the limits of what we can know about the history of art. I think that's really fascinating. And I haven't gotten to the point of, of who created what kinds of objects and what could we, you know, what I'm just, I'm still at the level of names. So that I think will be really interesting as well. Thank you. Yeah, and it is. It was really resonant for all of these parts to think about reading into the meaning of gaps and to giving those just as much as much importance as what is represented. I'm mindful of time. This is such a fascinating discussion. I see three hands raised and maybe so we can hear from Martin and then Paul and then Ming and then have to wrap up. Martin, do you want to go ahead. Yeah, I'll try to be quick because I've sort of addressed this elsewhere in writing, but something which I'm observing during today and thinking about a little bit is how, and this is really kind of the point which is the question that was being raised there. If we stop short of actually formulating a biography and accept that our identifications of, you know, the people that we're dealing with as artists or as art historians are always contingent and always qualified. If we don't kind of try and trace a life, but we actually stop short of that and say, well, we're going to deal with the residents of, you know, 10th Street or we're going to deal with the people who are at the Royal Academy, or we're going to deal with the people who identify as landscape painters and commercial directories. Then you then you do have a kind of data set, which is more reliable because you're not creating it, you're not creating it. There are kind of givens out there. Of course you can start mapping those givens on to one another and come up with certain suppositions or claims, but I think maybe that's an issue that we do kind of mobilise a set of expectations about consistency or coherence around which come with kind of biographical data. If we stop short of that, almost going to just work with the raw data rather than trying to form it into biographies, it might end up with something rather interesting. Thank you. Paul. Oh, you're still muted Paul. Very briefly I've been thinking about some of these issues and the idea of dealing with gaps by looking at contemporary data and learning from that and then projecting backwards with some of those lessons in mind. Because there's so much more data available more recently, it's like looking under a streetlight instead of out in the darkness. So there seems an interesting possibility to develop some strategies that way and also explore some different perspectives beyond the biography. Thank you. Ming, would you like to? Sure. I think that, so I did find it, I will put the website, the URLs in the chat right now, it's a website called Mapping Philippine Material Culture by Christina Juan and what's really interesting about it and the question that it asks for me is, who are these databases for? Who are we serving with these, the use of this technology? And what she has done is really interesting in that she has mapped Filipino objects of material culture in overseas collections, which then allows people in the Philippines to know where those objects are, to access what they look like, because in many cases, every example of those heritage objects exists outside of the Philippines. And so for craftspeople who are engaging in the process of cultural recovery, it's critical for them to be able to figure out where these objects are and what they look like. And many of them are not going to be able to travel to those collections, but she's able to share for those photographs with them, and also by speaking with the cultural producers on site, she's also creating a dialogue which provides greater knowledge about those objects. So I think it's really not just a question of what is the data that we're tracking, but how are we using it and who is it for? Yeah, thank you so much. That feels like a really great question to kind of end on in a way. The question of who is using these databases, projects, what audiences they are serving and for what purposes. And this has been such an interesting discussion. I'm really grateful that we have another space for discussion in the day so we can come back to some of these questions. And I think especially also around questions of artist biography and maybe alternative modes, whether it's looking at objects or thinking about where data around communities lives and what that looks like would be really fascinating. So just a huge thank you to all of our speakers on this panel. And really to everyone also for contributing to the discussion. It's been so interesting. And we all have a break now until 4pm. So we'll see you back then when we will hear from hands on this. Thank you so much everyone. Well welcome back everyone I can see panellist cameras going on so I hope those of you on screen as well as those of you who are joining us for this conference have had a very good quick short break filled with caffeine and other stimulating substances as needed to power us through to the, but the final part of what has been a really, really stimulating set of conversations and for this final section of the conference workshop whatever we want to call it. We've invited someone who we know is going to think with us through with a historical lens. And we've invited Dr Hans Honiz, who's a lecturer in art history at Aberdeen University to give a paper slash response slash thought piece that we thought would sort of not wrap things up but kind of presents us with some ideas and strands for further discussion. Let me just tell you a little bit more about Hans. In 21 to 22 he held the Paul Mellon Center's research collection fellowship, which centered on a project on British art historiography, particularly the collection of the papers of Paul Opay that we have at the moment. But Hans has also worked extensively on the history of art history and art theory since the 18th century, and has written and edited numerous books including those on Heinrich Wulflin and 18th century antiquarianism, and as well as projects on Abibarburg and his new monograph, which is a biography of Abibarburg, is forthcoming as well. And through our conversations with Hans, both around the research collections fellowship and other publications and other papers he's given for us at the Paul Mellon Center, we've really sort of opened up thinking about data gathering in different forms and mass data methodologies and their historiographies. And we've had really interesting conversations about data and disciplinary formation within the context of the Paul Mellon Center that's been thinking about the context of the formation of British art studies and the historiography of art history in Britain. And so those questions and those strands of thinking really made us think that Hans would be a wonderful person to speak at the end of today and give some historical or historiographical perspective to some of these questions about mass data methodologies. So Hans, thank you so much for being here and responding to our invitation and I'll hand over to you now. Thank you so much, Sarah, my greatest pleasure and I'm very much enjoying the day so far. Even though it is Sarah has already indicated quite a bit of a different cup of tea to what I'm normally doing. I hope you can see my screen in the meantime, if I don't hear anything to the contrary, I assume. Thank you. Sarah has already indicated that this is going to be more of a historical paper and I am certainly not somebody who can claim any active experience in working with digital methods in art history. And so I hope you will excuse the slightly left field of contribution here. And I hope you will also excuse that I think I have taken the topic a bit too literal maybe because what I am going to do is focus very much on the question of mass and the role of thinking about masses in art historiography. And what I want to do is looking back into mainly a 20th century German art historiography to think about the relation between an interest in data and an interest in masses as in larger populations. And I feel, I feel, I feel and fear that the relation between the two might not be quite as smooth as maybe maybe the title of the conference today does suggest. Right, this is enough of an introduction. I am going back to the foundations of art history as indicated. I mean, I am at liberty to do so because as all of you might know, this is a very prominent motif in digital approaches to art history as well as all of you will be very aware. A lot of writing about digital art history is indeed precisely drawing on the great names of the field in order to forge something like a genealogy, I guess, for current concerns and approaches. I am quoting him because he's also here today, Paul Jaskit, for example, in 2020 explicitly highlights that we should return to both lean to an art historian looking at vast numbers of examples of early modern European art, and that this makes us hopefully rethink what the very core substance of our field is. I see this importance of data in our writing of the early 20th century, and it's kind of epistemological status, if you want, is what I want to query in the following. I assume everybody knows, but essentially a principle, the point is that he wants to write an art history without names, so an art history that is not focused on the individual artist, but does a kind of distant looking exercise, essentially, where an analysis of artistic styles allows us to cluster artworks according to different art works. So, for example, clustering linear artworks versus painterly artworks, Renaissance art versus Baroque, German art versus Italian art, essentially what Paul's got here in the quote, just to reiterate that a little bit. I find it really interesting that Paul is in that quote, sorry Paul, again for dwelling on your work here quite extensively, I think you're absolutely right by the way in what you're writing here, so this is not meant to kind of start into a critique of anything, just using it as a handy quote. I find it quite interesting though that highlighted in the quote, are the opposition pairs of Renaissance and Baroque and Italian and German. One could say that these mainly mark two categories or two kind of clusters that are quite far away from the constellations that interest us primarily in this workshop. Ysbrydlen Edith has very little interest in categories of population statistics. He's not interested in things like artistic skills. He's not interested in workshops, and indeed he also very early on abandoned interest in connoisseurship and kind of clustering works around a single artistic name. Anything that has to do with people, in other words, seems to have not immediately a place in this art history with our names. And I think this is telling. Early attempts towards a big data art history, if we want to name it that, sorry. If we want to name it a big data art history, I argue precisely not interested in data about masses, about groups and movements as the workshop abstract phrases it. On the other hand, something like the mass was anathema to a man like wealthy. I'm giving you here a few quotes about from his early book Renaissance and Baroque published in 1888, where he speaks about the masses in Baroque art in a strikingly negative form. The masses of unformed blocks of stones instead of cornices, the cornice, everything burst out into the seams and chaos rains over the space. So where when Mars comes in as a category, it seems to be primarily associated with chaos. Baroque art, which is characterised by by massive tide, so by a massive nurse, if you want. Baroque art takes recourse he writes to a great just need strong modes of expressions that are caused by a general darling of the nerves. So those artists that have an affinity to Mars seem to be of actually quite a dubious mental makeup, if you want. This is a common characteristic in a lot of both things right. Mass is considered as exclusively negative. And that also spills over from the full description of what mass is into an analysis of social spheres. The chaos of the Baroque masses in both things is regularly compared to the aristocratic composure and order of the Renaissance. The Renaissance being an elite culture where a sect few are reining in complete order and distinct kind of separation from each other instead of the clustering that's associated with the term of Mars. So those things are history without names is anonymous is maybe big data. But I feel it is distinctly not mass data, not something that is interested in unified groups of populations. There's nothing in the theme of the socialist pathos of the masses, which is coming out at around the same time that he is writing. I'm showing you here just the color of one of the famous, most famous expressionists plays of the period. And Stona's piece of mass human mass a manj, where he kind of is one of these great pieces of literature that restages in a way to Greek chorus and brings that back into modern literature. And where the mass of the workers is one of the key features right that they appear as a unified home and speak with one voice in the chorus. And if if somebody has to be singing out and has to be speak on their own, then this person is called dynamic laws, the anonymous. So, and the idea here is that masses sing in unison. And that precisely is what somebody like Verflin clearly couldn't stand right. Verflin's approach and both is anonymous data driven idea of art history precludes precisely any such clustering according to population characteristics. Many respects, it literally voids the individual, and does not want to, you know, even give the individuality of a name to the protagonist of his art history. At the same time, however, Verflin was very much interested in biographical approaches, perhaps paradoxically so. He wrote famous books about Albrech Dürer, for example, who is time and again appearing as a special individual in Verflin's work. Yet the interesting thing about it is that he characterizes Dürer, not necessarily in a in a kind of affirmative positive way so with positive attributes that mainly in a negative way as a negative way of description. In fact, Dürer is not great because of what he did, but because of, as it's here in the quote, because of what he has overcome. The etching here, mryt y tawd yn taweth, it might be a bit stiff, but Verflin concludes it is this restraint, even for us today, that is the root of the moral power of this man. One is almost tempted to speak of a morality of vision in Dürer. What here is a description of Dürer, the artist, who is described as somebody who, well, quite similarly like an artist who without names, voids itself from himself from certain characteristics. His moral power is restrained, that takes back, pairs back the individuality that is often in biographical art history, seen as the driving force behind artistic development. And I find that fascinating because what I see here coming together is on the one hand side a data driven methodology that is interested in anonymity in an art history without names. And on the other hand side, a personal ethic that is also interested in well maybe not anonymity, but a pairing back of individuality. So the restraint to not come forward as an individual is here deemed as something that is evidence of a specific morality in an artist like Dürer. The decision to pursue data driven art history, in other words, I heard content, is in part motivated by a personal ethic that is driving Verflin personally. The ethos of an artist who have names is a form of denial that potentially can as a Dürer lead to a reformation of self. I hope what I mean here becomes a little bit clearer when I switch to another example, incidentally another man that I wrote a book about so I wonder always how you find your examples. But the man I'm talking about is Avi Warwick. I'm choosing him, of course, also because he too, like Verflin, is often seen as one of these kind of father figures of digital approaches to art history or more big data approaches to art history. I'm sure all of you know that better than me, that works like his Atlas Nemocino, where he kind of traces certain pictorial motifs across different periods and geographies that this often was seen as a kind of early stage of linking and networking pictures, to the social historical context, but because of something like, like, you know, what computer vision it does, right, recognizing motifs and then kind of following them across different periods. It could say a lot about what I think about this particular interpretation of Warwick's work. I shall only say here at this point that I think that it is important to highlight that Warwick was, while he was doing these experimental arrangements also a staunch defender of historical positivism. And that is the very reason why he was also very interested in data as a resource. He advocated throughout his life the accumulation of facts as a matter of personal academic ethos. Truffle Pig Services is what he once called that in his diary in 1907. The Truffle Pig, of course, is somebody who digs out facts, but then lets them go, right, and somebody else collects the truffle and eats them. He is not the one who exploits them. But this is precisely what Warwick earns the highest praise, because it is, as he once called it, a self-renunciatory labour that kind of does something purely for the service of a higher good. True science, Warwick wrote, is marked by calm and sovereign approaches and is free from all the grand, envious, grand delinquents. Again, for Warwick, working with big amounts of data is also a form of self-denial. As soon as you get into the role of staunch historical positivist who sifts through the archive and piles up materials, you are also the one who takes back your individual desire to shine if you want. Here in the quote on the top left in a letter to his friend André Iones, I think that comes through very, very clearly, where he says, I too was born in Plutonia, so the platonic realm of ideas, but such soaring movements are not for me. I have to look backwards to cast my eyes with a philological gaze down to the soil. So what is happening here is that a data-driven approach is very much contrasted with an interpretive one. Well, at the same time, the data-driven collection impulse is seen as the ethically superior one. Warwick himself indeed contended that the good God is in the detail, as he famously once noted, which is something that, not without coincidence, draws upon God and the language of theology in the expression here. By the way, there's many similar authors at around the same time to say things like that. Edward Sprengar famously said it's a task of the historian to study in the tangible singular things, the intangible divine home. What we've got here in general is, I think, a very strong ethical streak that is aligned with the demands of Protestant ethics that were propagated by leading theologians of the time. It's this idea, I went into that, that couture Protestantism, so Protestantism as a cultural force more than a religious one. Adolf von Hanag, for example, the most important theologian of his time, argued that self-denial is what Jesus demands, self-denial to the degree of self-renunciation. And I think what Warwick and Warwick both are advocating by taking themselves out of the equation, but by recoursing to the language of data, is to perform an act of self-renunciation, as here demanded in these Protestant ethics. So I think this is an important point. A tense data driven art history were frequently driven by a desire to abstain from more interpretive, authorial ways of writing art history. Statistics as a means of self-denial, if you want. An art history without names tried to transcend the merely social sphere and enter into a more abstract aesthetic realm. Again, I think this is just important briefly to highlight, not entirely, not restricted to German debates. The same impetus of the selflessness of data mining can equally be traced in British historiography. John Bagnott Burry, for example, the famous Cambridge historian of the early 20th century, or art historians such as Basil Tainor, the founding director of the Paul Mellon Foundation for British Arts, is also a precursor of today's institution who also argued for the importance of statistical research in art history and very much saw that, saw statistical research as an attempt to steer clear from projects and enterprises, which are really too ambitious for the present stage of knowledge. So, again, there's this idea that, you know, focus on data needs to quell over the ambitious freewheeling interpretation, and I argue that at least in the German debates, this is characterized by a self-denial and a Protestant ethos of renunciation. So data, in that sense, very much seems to me opposed to the idea of mass, opposed to the idea of an association of people who belong to a kind of amorphous social group. And I think this is corroborated when we look at some of the authors that quite soon after Welfin turned to this very category of mass, turned to categories of populations that are bigger than one, or at least not entirely anonymous. So, are we talking about mass versus data? I think we do. And in order to give more meat to the brain here, I would like to look at two of Welfin's successes on his chair in Munich. Welfin was a professor at Munich until in 1924. He took quite suddenly early retirement and left this very important chair vacant, and then he had a succession of successes. I will focus on the first two here today, and those are Max Houtman and Bill Heddon. And what I want to show you here is how their interest in masses in bigger amounts of populations, where on the one hand side an attempt to kind of push back against Welfin, to push back against this very renunciatory idea of anonymous art history without names, but also on the other hand side the attempt to still kind of do justice to Welfin's legacy of an quasi scientific history of art. I'm starting with Houtman, who's certainly not the very well-known character, also because he committed suicide actually after only, I think, two years after being in post in Munich, it was never quite clear why he got appointed in the first place, to be perfectly honest. He was mainly a historian of the very church architecture, and that is also the subject of his most ambitious work, the history of church architecture in the modern period, essentially. Published in 21, so six years after Welfin's principles. The interesting thing is that the preference on the one hand side makes very clear that this is a work that is committed to follow Welfin's principles and the kind of developmental ideas that he's presenting here. But on the other hand side, it is also a work that is not based on data in the strict sense of the word. It is not a catalogue style text based on archival work, etc. Instead, he wants to flesh out the great lines, the historical structures of development. For our purposes, I'm particularly interested here in the first chapter, which is dealing with architects and patrons for the very church architecture. Here, Houtman draws on an empirical statistical way to map artists. This is just one of the many charts that illustrate this book. What he is doing here is to map essentially the structure of organization in church architecture. So he maps, for example, the profession of individuals, he maps in which time they flourished and work, and he maps whether they are indigenous German artists or whether they are foreigners who came from somewhere else. So essentially a big table of what was going on in the very church architecture. He has a second of these diagrams that he is drawing, where he is mapping the indigenous German architects who increasingly become more important against the foreigners who are here to dotted line that remains a bit more unimportant. So, on the one hand side, this seems to be a very empirical approach, if you want, but I think it is anything but merely statistical. The aim of Houtmans is rather to filter out two trends. First, the gradual prevalence of trained architects and builders. And second, the rise of German artists who slowly sideline the foreigners that dominated Renaissance and early Baroque art. So once we come to the 18th century, the Germans, the indigenous people are much more important. There is no doubt about the contemporary thrust of the argument that is happening. In fact, much of what he wants to show us here is an aporia of the effects of war, for example. He writes about the 30 year war and how that essentially cleans the entire German population and puts an end to the unrootedness and hybrids who don't know where they belong. The war, he writes, leads all relationships.