 Thank you everybody for coming here. We'd like to talk to you about how you believe that Limetop and data could be achieved in the cultural sector by, you know, getting into the co-creation world. So first a little word about PACT. So we are the Center of Expertise for Digital Heritage, funded by the Flemish government. And we started out as an institution for archiving of audio-visual heritage, but slowly grew into an organization that talks about all processes related to digital heritage, such as cataloging, such as publishing your data, but also reusing your data. So and our center of concern is doing this in a sustainable way that allows people to access your data, regardless of your institution sustainability. So before I want to make an argument for opening up data, I want to kind of consider what the museum is and who it is meant to serve. And I think if we take a historical perspective, this is probably a good way to see how this shifted throughout time. So this is a picture that shows an imaginative reconstruction of the library in Alexandria. And this is probably the earliest at the station of what led to the modern museum. But in fact it's actually much closer to university in modern terms, since they do not necessarily take after a collection. I mean it was a library, but it's mostly a place that gathers scholars. So the other example historically of the roots of the museum would be the Cabinet of Curiosity from Renaissance, where you can see a very explicit focus on objects, but it's about showing all of the objects because the point is not necessarily scientific study of these objects, but to show them how the owner of these objects has an influence in the world. And of course the audience is mostly courtiers or privileged visitors. And then finally we arrive in the Enlightenment era, which has the roots of the truly modern museum in a certain way. But we're not quite there yet because in the beginning I guess it was also mostly about scholarship of objects, but putting them into a taxonomy, but not necessarily displaying them to large amounts of people. But this of course changed in the 19th century where museums would open them up to large middle class audiences, also as a way to bolster national identity. So here we really see this combination of having a research community, but also doing outreach. And that is still very much reflected in the modern ICOM definition of what a museum is supposed to be. But we're going to try to make an argument for how this could be achieved, opening up your collection, studying it in the digital era. Because basically we believe that consumer demands are changing, and that's true in the commercial world, but also in museums. And I have two kind of central ideas that I think kind of link into this. So the first thing would be on the right, that's a book written by Lawrence Lessig, which is called Remix, Making Heart and Commerce Thrive in the Hybrid Economy. It's mostly an argument for publishing public domain works under open licenses. But at the same time it also makes an argument for what he calls the ecosystem of reputation, which basically means that in the digital realm, museums are already in competition, which with quote-unquote unofficial voices that are writing about their collections that actually have a quite large audience. And I guess Wikipedia, which is what we're going to talk about later, is the prime example of that. And then apart from that, this graph represents the long-tailed. And this is also kind of an idea that in a digital world, basically shops that have no physical manifestations, such as for instance the Amazon Web Shop, they are able to sell a lot of things that are in very low demand. So things that are interesting for people with a niche interest, but altogether that creates a very large market. So if you look at the middle pie chart, so for Amazon apparently that's 75% of all their sales. And if you apply that to a museum content, we believe that apart from your best sellers, if we can extend the analogy, a true open data strategy would allow you to serve quite a large part of people that you are not giving the tools or the raw data that they would like to work with. So these are two central ideas that I think we want to argue for. So I talked about the ICOM definition earlier. So a museum is a non-profit, a permanent institution in the service of society and its development open to the public, which acquires, conserves, reach searches, communicates and exhibits the tangible and intangible heritage of humanity and its environment for the purposes of education, study and enjoyment. So we believe that researching and communicating is now still mostly interpreted as institutions controlling the quality of curated content that is made about their collections and also controlling access to their data. While we believe that in the future museums should profile themselves as data publishers who share their data, contextualize that data, but that also coordinate initiatives that do this and facilitate people who are interested in their data to do something with it externally from their institution. So my colleague Alina will tell you something about accumulation and registration of data. Hi, so yes, I will be talking about, we wanted to tell you about what is going on in the back office of the museums and some will talk later about publishing front office data. So about accumulation and registration of data. If I may make a generalization, it's not true for all institutions, but in overall for Flemish art museums, this is the situation. The back office technology and infrastructure is not a popular topic to invest in. It's not something that you can make quick wins, fancy apps. So it's often not thought of or not in the main vision of the institutions. There is often also no digital, sorry, digital mindset and often the IT services are external. So there's no IT personnel in the museum, so you're often very dependent on these external providers. The data is messy and not complete. We're talking about basic registration of art objects, but the whole context of it is just, there's just no time to do that. But there's a lot of knowledge in these institutions. The data is captured on different carriers, so in systems, but also still in analog paper dossiers on other Excel outside of the system. So it's a big chaos. They're often still used in closed and obsolete software. So when you ask questions about link-top and data, it's often very difficult to put it in the situation of this software. For audio-visual material, I'm talking about the reproduction, for example, of artworks. There's a lot of digital asset management systems in the market, all not good enough, and there is still an analog mindset of making a copy of the image for every project. You make a copy, you put it somewhere on one computer on the other computer, so you have a lot of different copies, and audio-visual material is not seen as a resource that you can always take for different projects. We at PACT had a chance to initiate a series of projects on optimizing this data in museums, and this is a project on a data hub that we did together with the Flemish Art Collection. It's a partnership between the Fine Arts Museums, KMSKA, Gruninger and MSKHENS. Some contemporary art museums also took part, and it was a project on analyzing the infrastructure that is now the outdated infrastructure and making a plan for the future for a better platform that can help museums to function as these online knowledge sources. And I will talk about the end goal of this project was making a data hub for these museums, which aggregates the data from the collection management systems and puts it via an API as open data for reuse. I wanted to talk about different projects around this data hub for you to see what's going on and what kind of open data principles we tried to implement in the cultural heritage sector. And so the first question was the persistent URIs, Persistent Identification of Artworks, which here is a link to an artwork from 2012 of SMAC, which of course doesn't work now because the museum has a new website, so if you're a researcher or I don't know who saved this link into your data set, you have a problem. What we did is looked for a way for this persistent URIs and persistent identification of artworks to be implemented by the museum workers themselves, so not again an external IT service, which you have to call every time for every persistent URI to be made, but is over a tool that can be used by the collection managers themselves, which we developed ourselves. It's an open source tool, the code is on GitHub, and now for example you have a persistent URI for an artwork that resolves to the, for now just the HTML page about an artwork, but in the future you can put whatever you want behind it. But the problem is that this resolver tool is still the persistent URIs, so concept that is too hard to process for collection managers, and this tool needs to be developed further for it to be a success in cultural heritage. We also played around with enrichment of data via external authorities, so you talked about taking data from external authorities and adding it to your data, which was very interesting to see how all of this, I think it's the example of Peter Brögel, that has different kinds of his name and notations, which you don't have to copy paste anymore as a collection manager, but which you can just take from other authorities. All this stuff, we also tested how you can internally reuse all this linked data from different museums, if you aggregate them in one data hub, for example visualizations like the acquisition source, like where do artworks come from into a museum, and you could see all of this artworks from the church, artworks from the artist, artworks from the king that were given to a museum, so fun stuff to see also just for the internal business process of the museums. The beta version is also on GitHub, so it is now tested by museums. The data is being aggregated from the different management systems, so we are looking forward to the feedback of the museums. For the other visual material, my colleagues looked into the situation on reproductions and made a blueprint for a distributed management of this material using IIIF and these kind of open principles. All this project made it possible for us to start testing the linked open data publication for the cultural sector and this is where some... Basically, PACT started with open data publication with a website called Open Culture Data in 2012 and it was not very successful in the sense that it's data dumps which have outdated information and also the developer community didn't really find their way to the platform, so that's obviously I think something I heard a lot of people talk about today. Basically, PACT started looking elsewhere and you are probably all... I don't know if I have to explain, but I'll say everybody is basically aware of Wikipedia, the encyclopedia, but apart from that, of course, there's a Wikimedia ecosystem of platforms that work together. Wikidata, the database, Wikimedia Commons, the audio-visual repository which provided us a possibility to start opening up collections and why did we choose for Wikimedia? First of all, it's a living system, so not only will people reuse your data, they will also engage with it so they might add data, they might link it to other sources and also, of course, very much to the country or the website, it has much larger visibility, so people will find their way to the data that you publish. So as Alina already explained, since persistent identification is a basic condition for linked open data publication, those persistent identifications are already there for a number of fine arts museums in Belgium. So the second step was to get them to publish a very limited amount of data fields such as a title, creator, some dating information, for instance, and together 27,000 works had their data published on Wikidata. So then the second step was how do we get images of reproductions for that because as many of you might know, the first decade of digitization was done according to a business model which made organizations claim a copyright on their images. So first and foremost, we started doing a crowdsourcing project with Wikimedia of Belgium which is called Wikiluf's Art, so this is still from it, where we would go to museums to ask them if we could go on a tour, if we could do, for instance, a real photography setup to make reproductions of their artwork. And I go mad if Wikimedia of Belgium is right there, so I help them a lot in the process. So this was an interesting format because, one, it didn't require the museum to change its right management statements immediately and it also didn't require them to give their reproductions. But, of course, these are not always the kind of images that have the same quality as, like, of course, a reproduction that's made in the studio in high resolution with the right color settings. So basically we are right now asking these museums to give us reproductions of a decent resolution and the coupling of the Wikidata information to images has already been done for the museum for 842 works, so that's a finished project that we're happy of. And we're trying to get all these other museums to also see whether they can do this, which is of course not always simple because of course there's also smug contemporary art museum in there. So what happens once organizations publish their data? So what's interesting is there are already a lot of very nice tools that are created by the Wikimedia community. It's a little vague here, but for instance right there you can see that this query is basically for this HUNI museum set that we just talked about. How many views these images got altogether for the month of January 2018? And as you can see below that is 44,987 views, which is quite impressive. So we're trying to make institutions to think about that kind of visibility, apart from what they are already used to. And furthermore these tools give you interesting information, for instance, in which articles the content is used. So that's not necessarily only an article about the specific painting, but it can also be a theme that's related to it for the French Revolution for instance, or a movement. So it can be kind of reapplied in many different ways. And you can even check when it gets peaked, because for instance if there's a temporary exhibition going on, obviously all the people are looking for that data. And yeah, you see a lot of interesting curating projects that are done by the volunteer community. For instance, this is a fur fashion in 1562. So we just noticed that one of the images that we uploaded was of a man. Yeah, it's over here, it's a little vague, but it's basically a painting of somebody with a medieval fur coat. And this is put into a collection that is curated on Wikimedia Commons about fur fashion throughout the century. It starts in the third century AD and it goes all the way up to the present day. And this is perhaps not something a museum would work on because they are concerned with their collection, but nonetheless it's a very fun example I think of something that happened almost immediately after we published the image. And apart from that, obviously content, data and images on the Wikimedia ecosystem does very well in Google rankings. So apart from the fact that it usually shows up very high in Wikimedia Commons, you have the Google search box which draws content from Wikimedia, but also images from Wikimedia Commons and some data over there. So it provides a very nice visibility. What we also noticed is that data does get enriched. So for instance, this data about exhibition history, this is not something we added, so that was added by somebody externally. And as you can see, European 280 there is referenced, so it's high quality data. And underneath you see that certain iconographical information is actually referenced using standardized codes. So it's the data that the heritage sector wants to work with. But the problem right now is how do you harvest this data to get it back to the data of the institutions? At the moment there is no automatized way and certainly not a way that a collections registrator with no particular IT background can just get it out of these data at least not on an automatized way. And we also, we didn't notice that some data definitely travels, so for instance this is the virtual international authority file, so it's information about creators of artwork for instance, or persons. And this is the record of James Ensur, and obviously that was one of the data sets that was released on certain creator and it's now used by VL, so they harvest our data. And furthermore there are all kinds of neat tools that are already there that you can use to visualize your collections. So obviously if you have reproductions images and data you can do all sorts of night stuff like bubble charts and timelines obviously. You can map your entire collection on a timeline like this for instance without you really having to write software or create something new. And this for instance is a tool created that's called CODOS. And basically if you do have your collection released with images you're able to automatically generate a catalog for your collections. So we would like to see more of those tools that are easy to use and that are interesting for the cultural sector to be created in the future. Let's go to the conclusions so we can go into that. Okay, so short the conclusions and the next steps. Why were we talking about linked open data in Limbo? It's because the institutions, the museums still struggle about this concept and it's often too extreme. Yes, between two extremes and the one is wanting to participate in this digital movement but having no digital mindset or no vision on how to do it. They want to reduce costs on IT apps but it often also can end up in an out of control IT budget if you don't do it sustainable. So you keep on adding stuff, you keep on the collection management systems adding stuff and it just goes out of control. You have the new digital audience that are good with smartphones but you also have an audience that still doesn't use any digital software and apps. So how do you interest both of these audiences? There is also a very hard pressure budget-wise also to show off with fancy tools and let's know that you're participating but often get locked up in absolute technology so you keep on creating new applications and you want to engage with the crowd but then you have a mass of abandoned web portals that nobody ever uses because nobody knows they're there. So this is kind of the Limbo situation. Our wish list would be for ASAP more IT profiles in the cultural heritage sector because we really see that if somebody with an open Belgium mindset comes into the institutions it really goes faster. This one is also talked about today. There should be more freedom to play around and to test things out. The budgets that museums get now is just you have one try and it should be good. If you fail then you don't lie about it, you don't talk about it but you should get the freedom to play around because the tools that there are now are often should be made or custom for institutions and there's not like one resolving service for everything so playing around is very important. Better open infrastructures and tools so for back office the collection management systems that are being used now are very often closed and obsolete. More tools that give the possibility to publish open data online. More examples of reusing like Wikimedia data reusing and stuff like that and I don't know if you see that more vocal and demanding audience because we are often find ourselves in the loop of yes but nobody is asking open data so why should we do it? And then the reusing community says yes there is no open data so we don't have examples of how we want to reuse it. So this is what we need to break out of. PACT for our future plans is to work further on this data hub project. We are giving our resolver tool to the open summer of code in the hope that something better will come out of it, a better usable tool. We have a set of Wikimedia projects like about publishing artworks that have come into public domain yearly like structural publishing of data and images. Also reusing this information in Wikimedia articles but also outside of the platform. We are also looking for the current, we wanted to do a survey on the current demand of cultural heritage data by various sectors so not only the traditional tourism but outside of it the research digital humanities communities etc. So if you see yourself contributing to any of these questions or the wish list that you have seen do you have ideas or feedback please contact us we are very eager to hear it. So thank you, do you have any questions? I mean why are you trying to use your cell phone to develop your museums? Because it seems to me that museums in different countries especially in the new exchange are supported all the time. So would you expand for other countries and try to create a regional part of museum data? Well there is already, of course, the European, this is the big aggregator of the data from all European cultural heritage institutions but what we have seen after the years of this project is it's also it's very hard to build this one big aggregator. For example nobody thought of asking for persistence to arise to reproductions when they took data in so now a lot of museums are just showing like the error with the picture because so now they have to do it again. The data hub which is built in this project is I think the future is behind smaller data hubs which then go higher and higher because if you create one big thing then it is hard to make it sustainable to control it to let it be accessible on long term. So and the code is open source I think it is a possibility that more museums could go into this data hub but of course also one of the problems that we are working with it's because it's project budget. So you have one project and you need to fit it in some kind of which is really a risk for sustainability of this kind of project but we hope by doing it in open source that it will create a community around it. Quick question. Does it have a sparkle at the point of data? Can we interface with it? I think it's the OIE. I think it's the you have to be with my other colleagues for this kind of questions but it is the possibility there is an API which you should get possible to get that out of it but I don't think it will be sparkle. I think it's the data that is in there is Lido XML and I think to get it out So there are no triples? Yes I think it is like a triple data but you really should be with my other colleagues I'm sorry. To answer maybe part of the question wiki data query is actually sparkle. If I want to carry another museum and link it to ours I have to go through wiki data which is absurd. Also maybe it wasn't clear from our presentation the data hub and wiki data tests are next to each other So you have the data hub where more data will be available the wiki data was a test to show how you can also use platforms, free platforms that are already available. Maybe be clear on the wiki media ecosystem it's a way to display and open up your collection but it can never serve as the same service. Yes as your own online source because also everything that's on wiki data should be referenced so you cannot create, you cannot put original content on these platforms so it's always important that your own source online keeps existing as a source of the original information.