 All right, I guess we'll go ahead and get started. Welcome to this morning's panel. This will be a true panel. We're not going to give any talk. So I want it to be conversational. There will be time for audience questions after we run through our questions. And so please, if you do have questions for any of the individual panelists, just hold them till the end, and then I can run around with a mic. We are being recorded, so I don't want to have to interact with, break things up to get a mic out to you during the presentations. So my name is Harrison Decker. I'm an associate professor and data librarian at the University of Rhode Island, and I'm the director of the library AI lab. I'm going to introduce our panelists and then just get started right away with asking them questions. So the gist of this morning's panel is that I wanted to address. And this is sort of prompted by a comment someone made to me who works at a library that is fairly heavily invested and offers great and extensive programming training in languages like R and Python and a variety of other, and GIS and a variety of other techniques. And they said, despite the reputation, despite everything they're doing and all their success stories, they still get questions from the university provost about why is the library actually doing this. So that got me thinking, well, maybe this C&I would be a good venue to have a conversation by four practitioners who are very heavily involved in just doing this sort of thing. So to my right, I'll start with our first panelist, and he's Matt Burton. He's Matt as a lecturer in the Department of Information Culture and Data Stewardship at the University of Pittsburgh School of Computing and Information. His research focuses on digital humanities, scholarly communication, and computing education. He teaches design, data science, and web technologies in the master of library and information science program. And he has a PhD from the School of Information at University of Minnesota. Sorry. Let's make that University of Michigan. Yeah, yeah, yeah, yeah. He does have a PhD, though. OK. Seated to his right is Vicki Steves. Vicki is a librarian for research data management and reproducibility, a dual appointment between the NYU Division of Libraries and the Center for Data Science, and her research centers on integrating reproducible practices into the research workflow, advocating openness in all facets of research and building, contributing to open infrastructure. And she is the co-founder of the LIS Scholarship Archive and Open Source Open Access Repository for LIS and Alive Fields. She also works on the ReproZip project, an open source tool that enables full computational reproducibility of research and TAGET, a free and open source qualitative analysis tool. Then we have my colleague from URI, Indrani Mandel. She is a lecturer for the Department of Computer Science and Statistics of the University of Rhode Island. And she is the education coordinator for the URI Library AI Lab. The Library AI Lab is only a year old, but has already been in the headlines for its undergraduate research and outreach programs. And Indrani's interests are machine learning, data science, and AI education for K-12 students. And then our final panelist is Tim Dennis. Tim is the director of the Data Science Center, a research and education support unit of the UCLA Library. He is an instructor and instructor trainer with the carpentries, an instructor community that teaches foundational coding and data science skills to researchers worldwide. He recently worked to establish library carpentry as a lesson organization within carpentries and serves on its curriculum advisory committee. Tim's interests include research transparency and reproducibility, open data, and inclusive pedagogy. So to get started, I'm just going to go across the panel and we'll just, I guess, start. We'll start at the far end and work our way this way. I just want you to briefly talk about your own educational background and the work you've been involved with and what sort of specific programming topics you've taught or computation-related topics you've taught. So start with Tim and work our way down. So my educational, I guess, background, I have a master's of information management and systems from UC Berkeley, which is called iSchool now, I think. And it's part of the Data Science Division, which is relatively new there as well. And when I consider myself really a data librarian and I'm self-taught as far as programming, I learned a lot at work with Harrison, or moderator at UC Berkeley, and learned a lot from him since he has a programming background. And at Berkeley, I learned a lot because there was this data science kind of renaissance that happened there as part of this Sloanmore funding for data science environments. And it was really highly generative. And I've been taking what I've learned there and applying it to other places like UCSD and now UCLA. The programming topics I've taught are Python, Git, Bash. I've taught data visualization and those tools. We've recently, I've taught text data mining and R. So we try to cover a wide gamut of things and I'm actively involved in developing the Carpentries Program at UCLA. We're a member of the Carpentries. It's like a member organization like CNI itself. And we're developing, you know, more trying to scale up our number of instructors on campus and teach more of the workshops. So I think that's kind of. So my journey started in India. I am an engineer by training. And I'm an electronics engineer. When I moved to the United States, I joined University of Rhode Island and that's where I went to grad school. That's when I switched to computer science and got interested in machine learning and data science. Currently, I'm working there as a lecturer and I teach computer science, core computer science courses. Additionally, I also teach programming languages at the library. I teach R, Python, machine learning. And I also work with undergraduate students interested in research at the AI lab, which is located in the library, which is a nice thing because we get students from very diverse backgrounds. As for me, I also came into my higher education thinking that English or the classics would be the best thing to do before library school. And shout out to Nanette Bayer, my old advisor, was like, oh, you should actually do computer science and just try it, see if you like it. And so I was actually sort of coerced into starting this journey that way, although very happily. So that was where a lot of my background came in with learning some of the basics. And then when I started, I actually thought I was going to be a corporate librarian. And that was totally disrupted again by NDSR, the National Digital Structure Presidency, where I got a really deep dive into working with different types of data and computational processes. And that got me really interested in the way that we are teaching new researchers how to conduct their work. So as a part of my role, I teach all of the programming languages and concepts that my colleagues here teach. I would say one of the things that I am also teaching are things like how to ask questions of data for machine learning or AI purposes. I have a lot of consultations with folks who are like, how do I make sure that when I'm fitting some data for a model that I'm not completely biased? It's a hard thing to answer. So alongside of the computational literacy stuff that I think is really important, we're also seeing a niche data literacy on the rise as a part of supporting data science as well. All right, so I have a very long circuitous journey, so I'll try to keep it concise. I have an undergraduate Bachelor of Arts in bioinformatics. And so I have some formal computing training, but also situated in a science context of scientific computing. And then I worked in industry for a while doing information security, and actually wasn't really on a path towards libraries at all. And then one of my office mates was like, this information, we were talking about the challenge of information security. And he's like, this is really actually a library science problem. And I kind of pivoted and ended up going to the University of Michigan at the School of Information, which kind of introduced me to the information schools world, where I am today. I ended up getting a PhD in information and doing a lot of work around. A lot of my dissertation work was in kind of science and technology studies, and actually not computational at all. I didn't even tell people I knew how to code initially. Because one of the things that happens is if you know how to code as a grad student, you very quickly get grabbed by a thesis advisor because they desperately need people who can code because they don't know how or they want people to do that work. And I saw that happen over and over again. And I've seen that happen. One of the things I started doing at Michigan is trying to address this issue. So I started working with some faculty on developing kind of a computing education workshop, kind of similar to the Carpentries, although I developed a little bit independently. And then I did kind of get involved with some of the Carpentries movement. And then since moving, then I moved to the University of Pittsburgh. Initially, I was actually jointly appointed between the library and the school. At the time, it was the School of Information Science. I helped launch a digital scholarship unit. And we actually created a workshop series that was a joint workshop series between the School of Information Science or what was the School of Information Science and the libraries on kind of running workshops on technical topics like web scraping and data visualization formed the School of Computing and Information. I moved full time, so I'm teaching faculty there. And I teach a bunch of different computing. Computing is, I teach a bunch of different programs. So I teach in the undergraduate computer science program. I teach in the Masters of Information Science. So I teach kind of big ideas in computing and information, which is kind of a conceptual course for the undergrads, but also has a skills component to it, which is like teaching undergraduates, how to do Python and Command Line and Git. And this is computer science undergraduates who actually never learn that. They're just kind of expected to know it by the faculty. So they're like, well, why should we teach them the Command Line? They should already know how to do that, which is a very non-inclusive approach to some of these kind of very basic skills, which I think is that attitude and that need for the kind of addressing the basic skills in a semi-formal setting and for all of the students as a requirements or as a necessary step. So I've also developed and worked on this research computing education initiative that is running an eight-part workshop series called Data Basics, and that's actually for graduate students in the disciplines who don't have an opportunity to learn, because the basics, so we teach Python and Pandas, which is a data analysis tool. We teach Jupyter notebooks and the Command Line and Git. So kind of a lot of the similar topics to the software carpentry in this kind of eight-part series that's open and it's free and open to anyone in the university who wants to take it, the graduate students, faculty, and staff. And the goal here is to kind of give, it's mainly graduate students, doctoral students, master students in the disciplines and sciences and social sciences and some of the humanities too, to kind of learn those basic things that they wouldn't necessarily have an opportunity to learn because it's not taught in formal coursework. So I teach that and then the other final piece on computing education that we are working on as we've, at Pitt, we've recently redesigned our Masters of Library and Information Science and one of the core components of this redesign that we're still working on is putting and embedding computing education for the library science students. And we're still, this is going to be a multi-year journey but one of the things that I want to see, the library science program is kind of an alternative path to learning computing for people for whom the path, the traditional path in computer science or in information science, that door gets closed or they feel that they're not included or are able to participate in those communities which is a recurring issue. So we're trying to kind of, and one of the key components for the library science program, I haven't really figured out, we haven't figured out how to do this but there's a huge demand for computing education and computing needs broadly in higher education, in K through 12 and generally. And so part of what we're trying to do is not only teach the library science students computing topics but also equip them and teach them how to teach. So kind of a train the trainers thing because Pew had a study that people are looking to their librarians to help them with computing education topics, especially in public libraries. So if we can kind of equip the students with these skills then they can kind of go forth and help address these broader needs. So to kind of jump right into the, to ensure that we actually address the topic of the panel, I would like to just toss out the question to the panel. From your experience, what are the pros and or cons of teaching these types of topics in a library? Just quick thoughts on that. And I don't know, I'll just, someone want to start picking. Oh, you're looking right at me, okay. We'll start with Vicky, but they'll never just jump in if you have something to say. Yeah, so teaching data science topics in the libraries, the pros and the cons. So the pros I think that Adrani was mentioning is that people feel like the library is a space where they can come and fail if they need to. So a lot of folks who are coming saying, I don't really know the basics of AI or machine learning topics or predictive learning, whatever, natural language processing, text mining. I don't want to go to this workshop in the Center for Data Science because they're going to assume that I know every, like all the toolkits walking in and I really don't. So the library becomes a space where they can try and fail at these new topics, which is a key part of the learning process. So I think there's sort of a safety net in knowing that I'm here to provide you the service and I'm not judging your computational ability. And in fact, I'm here to scaffold you so that you can reach that. One of the cons is like the opposite where people are like, why does the library know or care about this? Which is what prompted the panel. So like how would walking into the library and seeing me, like they don't equate that with data science expertise for lots of reasons. Being in a library, being a woman. So that sort of cultural problem is yet to be addressed. I think across the board for computing education, but especially in data science where the demographics are so homogenous and badly skewed. I would like to add a few more things. When we run our data science or machine learning or natural language processing workshops, we have professors, graduate students and undergraduate students all taking the workshop at the same time. All of them have similar questions. Again, what Vicky has mentioned is a safe place. Everyone is treated equally and they feel that I'm not an expert. I'm here to learn something and it's okay to ask any kind of questions. We, including myself, when I go to a workshop, I don't feel like I'm a faculty. I feel like I'm going there as a student to learn something. And it's a great opportunity because the libraries are really providing this neutral ground where we can all go as students, as someone who is interested in learning a topic. I can add too, I concur that, I remember I saw a talk by a Fernando Perez, who's kind of the guy who started Jupyter Notebooks and he had a big slide, why they chose the library at Berkeley to put the Data Science Center Data Science Initiative in. He says the library is like Switzerland. So it has like a neutrality, like Gianni was saying, that is more welcoming. It also, I think for me, there's a virtuous cycle of, so data services have been part of libraries for a long time. Now it's, the nihilism is data science, right? But it's been doing a lot of this service and consulting for a long time in libraries. And there's a virtuous cycle that happens when you consult with people over time, you see all these use cases across the disciplines. Oh, people are doing text data mining across many different departments. So those use cases kind of can formulate and become a curriculum, right? So you know what to teach. People are struggling getting beyond their laptop if they open a data set with eight million rows and it doesn't open. So what are the kind of heuristics with that? When do they need to move to kind of cluster computing? When do they need to do some refactoring of their code? So I think that's a rich ground bed for addressing common problems in academia and then you can develop lessons and teach and include a diverse set of people that are across disciplinary. So I think that's a real strength of libraries. I think there's a credibility issue that might be the con. So like who, how do, you know, I don't have a CS degree. So why should anybody listen to me when I'm telling them how to code or whatever? But I think it through kind of a reflective practice when you work with people, they, you develop a good word of mouth and they'll come back to you. So I think libraries are a great place for teaching this stuff. One of the things that, you know, so I'm mostly teaching, you know, I teach and I'm a faculty member so my like incentive structure and my organizational life is kind of governed around terms and teaching classes and the needs, however, don't necessarily conform to that structure. And so one of the things I'm trying to do is collaborate and work with the library and also work with our high performance computing center and try to like kind of create this triangulation of people with, you know, the HPC as a computing infrastructure, the School of Computing and Information, we have some of the expertise and the library has kind of physical infrastructure and also temporal, a different temporal rhythm that I think, because one of the things is after the classes and there's always that, okay, I need, you know, so I loved your workshop, how do I do Python package management? And it's like, you know, and how do I use this or how do I install this specific package for this because I'm a biologist or I'm a molecular and biophysics graduate student which I just had a whole bunch of them in my workshop. And I don't have the ability to kind of do that support work, just there's not enough hours in the day. And so, but that's where I think the library can help because I think the key component of this is to build a learning community within the university. And I think that's the university, you know, the library is, that's what they're structurally set up to do. And so they can kind of, the hard part though, the con, you know, that's the benefit and the power and the potential, the hard part though is just trying to figure out the organizational configuration, like, you know, who's gonna pay for whose time, how are they, and library to their credit spends a ton of time thinking about the organizational sustainability of a service to the, she ran some of the grad students and faculty who were trying to work on this and they just wanna do it on the librarians. They're like, well, slow down, let's think about it. So there's a balance there, but I think when it, you know, when it does come, when everything can come together, I think it can be extremely powerful. So to pick up on that idea of learning communities, one of the, I think one of the limitations of this mode of instruction that we were kind of talking about for the most part, being the workshop format, is inherently problematic with these technical topics that, so because you don't become a Python programmer in a two-hour workshop. So I would like to ask, like a very practical follow-up question, and that would be, what have you done to address some of the limitations of the workshop format? For example, dealing with prerequisites, scaffolding, finding times. Just scheduling is very hard in the university setting because everybody has their regular course schedule. Students might wanna take something at night, but grad students and faculty might have outside lives that require them to go home. So can you talk a little bit? I wanna go home. Did I say that wrong? Anyway, so do you wanna go home now? So anyway, what have you done to address some of these limitations? I have taught the workshops in the evening. Actually the grad students really want it in the evenings or some of them do and then some of them can't. So I teach, sometimes I do the workshop series during the day and sometimes in the evenings. It kind of depends on if I have other evening classes. In terms of scaffolding, the workshop is very much a basics. And up front I tell people, I mean we've extended it so it's like eight, three hour sessions so that's a little bit of a longer engagement. So I kind of do set expectations up front. One is I don't have any expectation that you have experience with this. I have everything set up. They just need to show up with a laptop that has a web browser and then everything is running in a Jupyter Hub in the cloud. All the materials are on Git and automatically get loaded. So that's kind of a technical answer to some of this, which is they don't have, there's not this like getting everything installed and set up stuff so I kind of scaffold them in that way. But I do set expectations on you're not gonna leave this workshop like an expert Python programmer. One of the benefits I have is the students are super motivated because they're opting in to take this. This isn't a required class per se. It's an open free workshop so they are personally motivated which is not always the case in a teaching context but I do have that benefit so they work on that. But again I think the next step is how to, how to give them the ability to continue and I haven't really figured that out. I've had conversations because one of the things is like faculty will be like, we want all of our doctoral students to take your workshop and it's like, okay, that's great, they do it and then what changes are you making in your curriculum to then reinforce what they learn? And they're like, well, we can't do that. Our curriculum is that, we can't change it. So it's like well then it's not gonna work. I can't do this, there has to be substantive change in the disciplinary curriculum and classes have to kind of, okay, they'll have basic Python but you need to teach them the methodologies because I'm not teaching them the methods and you need to infuse them with the computing that they, you can kind of assume that they have and reinforce it so. Yeah, well I wrote down Jupiter Hub and Expectation Management so pull that thread a little. The thing I would add to that is in dealing with the workshop format, something that's really important to me is to build out open educational resources that they can always come back to after the fact. So all of my sort of data science focused classes are applications based. So I'm not, I don't teach like this is what a variable is in Python. I say like this is what a variable is and you're gonna put your part of your data set into it and so everything that I teach is, has the pedagogical underpinning of like you're gonna, you could modify this code lightly with a different data set and be able to work with it and that relies really heavily on me creating a larger scale open educational resource to point them to after the fact to walk them through some of the other projects. I would also say distributing the actual teaching like I collaborate with some assistant faculty at the Center for Data Science who come in and teach some of these intro workshops as well and we co-teach a lot of them and I think that also signals to the students something that there's a lot of support for in the university so they could come to me, they could come to this faculty member and I think that's really powerful and empowering to them when they're in the middle of like an intro ML workshop and they don't understand probability and et cetera. I have noticed running the same workshop twice is often at different times of the week is beneficial to a lot of students because scheduling is a problem. When I have the time, the graduate students might not have the time or the undergraduates might not have the time at the, so running it twice once to maybe during the work hours one maybe after five o'clock is a good idea or running it on Fridays between two and five has worked well for with our undergrads for me as far as prerequisites are concerned I have it uploaded on GitHub so they have access to previous lessons. I always like to teach them in a series of workshop that means you don't walk out of the workshop after two hours feeling like you are an expert so you have to keep coming back. That is an expectation I said in the beginning of the workshops and I have had great success with students returning for the entire series. Yeah, so there's a number of different things you can do to help with pre-wex or tech setup. You can have install fests, I've done that before. Like in libraries I like to get our systems people involved if they can become helpers at your workshop get people set up and configured. I taught an African American Studies class that's on the mass incarceration phenomena this last quarter and because it's for undergrads I didn't want to have all the setup stuff and I taught it completely in our studio cloud which is a cloud based R tool and that worked pretty well because I could distribute the notebooks to them and they could work through it with me. I always teach live coding style and I have a bunch of exercises within the class so people have points where they have to work through a challenge or stretch themselves or apply what they've learned. I think that's very effective. The last thing I've done is partner with specific departments. So we have a good relationship with the masters of social science and the masters are urban planning and we teach these workshops because you want to encourage peer learning and peer teaching. So if you get people to come from the same department together and you reserve some seats for them then you can kind of get that happening because they know each other and they feel more welcome, they build confidence and then lastly on the feedback loop you want to have them come seek you out for consulting or to get help because they're invariably going to want to do their research and then they've learned some basics in R or Python but then what is the next step? Oh I want to apply this. I'm having problem with this package in R and then they'll come get help from me. So you want to see it as a whole cycle. Teach a workshop, you get consulting, teach another workshop, you do it, try to encourage peerage and that works kind of well. Can I add one thing super briefly is that I always show how to debug in my classes. Like I will often intentionally put errors into my code to be like, oh no I have no idea what happened, let's go to the internet and find a solution and I think that approach also really helps students after the workshop to be like oh, I saw someone like who is my instructor run into a problem and solve it and follow the same steps and they can probably figure a lot out after the fact too. Yeah that's an excellent. I mean basically we are, like in the carpentries do now encourage people if you're really good typer which I'm not or a speller which I'm not. So I have error all the time but if you are that we encourage people to introduce error because getting over error, the fear of having errors is the real big thing and having kind of persistence and resiliency with that is something you really want to come out of a workshop. So definitely. So let's shift gears a little bit because I suspect some people in the audience may say these sound like great things but we don't have the staff to do that. So let's talk about the staffing side for a little. So one is I mean we could probably spend a lot of time talking about what's missing from the MLS curriculum. So I don't really want to go down that path although if anyone wants to bring up any ideas or if you know of new curriculum that's being introduced but since we have a couple of people here that are doing things kind of on the margins of libraries or between in the intersection of libraries and computer science, maybe and I'll address this to Andrani first, talk specifically about what libraries and librarians could do to attract more computer and data scientists like you to come and work with us. I think the initiative needs to come from above like the administrators provosting to encourage this environment of learning and data science and computer science and encourage libraries to set up this kind of units or spaces, providing I was at another talk yesterday and they just created this space which attracted lots of students which would not have happened if that space did not exist. So URI has been proactive with it and our provost and libraries dean who is here has been very supportive. I am among the three faculty who are associated with computer science and libraries. What would attract us is providing us this environment and this opportunity to work with diverse set of students. It really helps us. We don't usually have access to this diverse set of students when we are tucked away in the computer science department or in the engineering department. Library really helps us to find the students and this diverse projects. We have had great success over students from nutrition, biology, social sciences to name a few. Yeah, so it just provides us with this incredible amount of wealth where the students are our resources and where we can be helpful and useful to the community. Matt, would you like to add anything to that? So the question was making computer scientists. This is something that we are very, yeah, so actually when I first started at the University of Pittsburgh, I was in a jointly appointed position across the library and in the School of Information Science as kind of an explicit effort that's similar to kind of connect the department, the academic department with the unit, the library unit. And there's a couple of us and that's kind of we worked on developing the Digital Scholarship Services unit at the library. We don't currently have anyone in that position but I think actually I'm really interested in these kind of joint appointments and kind of creating that space. I mean, this is something that we are trying to do. I mean, that relationship is also working itself out in the School of Computing and Information because we're computer science, information science and library science all together in one academic unit and we're all very different and so we're all kind of sorting out and building up those relationships. And this is a new school that's just been recently formed in the last couple of years. We are also hiring teaching stream faculty, tenure stream faculty. We are looking for graduate students. So if you know, if you're, yeah. Anyway, come see me, we are, because like the, we're doing this, I'm all this stuff I'm kind of doing on my own. On a very lean, I get one course released kind of work on all these workshops and stuff. I get somebody, and a lot of it is just personal connections. I have somebody from the high performance computing unit who is a friend of mine who like helped me spin up the Jupiter instance. And then I have connections and people in the library who are, and it's like a lot of it is like personal favors and just personal relationships. And I think now it's, now we're at this place of like how do we institutionalize this? And that's kind of where we're at right now and taking that, I'm really good at like finding the people who are on the ground who are gonna do the work, but now it's got to socialize up to the administrators to kind of get them to do buy-in. I forgot to mention another thing. We also have Joan Peckham here with us who is a, would you please raise your hand? She is a professor in computer science department and also works with us at the library. Yeah, it's another approach you can, it at least has been effective for me is that, and you know UCLA is a big library. So we have like a lot of programmers, one of which has a master's in machine learning and computer science who works in our software development team and so he's become a carpentries instructor. So he teaches with us and he also consults. And I have in my unit have two programmers who are now kind of being refashioned into data science specialist. And part of that is we provide training and support for that. So look within, if you have a big organization, look within to kind of fill some of these roles. And if you have a digital library like we do that are doing kind of machine learning on their collections, that's a perfect way you can kind of externalize that as a service and get those folks to help consult for your unit or build a unit, right? So this can be research facing and not just internally facing. So that's what we're trying to do. Okay, I wanna, I know we've been talking about the carpentries a lot and I wanna make sure we talk a little bit more about that in case people aren't familiar with it. I can't remember when I first learned about the carpentries but I think it's safe to say as probably one of the early librarians to go through instructor training six or seven years ago. And at the time I, this was at UC Davis. And at the time there was one other librarian involved and she was unique in that. She was a former research scientist with a PhD, Amy Hodge from Stanford. And I thought, you know, this is great. And this is like, this would be ideal for librarians but the cynical part of me thought ain't gonna happen. You know, librarians are busy, they're skeptical about some of this stuff but then they proved me wrong and now it's really taken off. So many, I think I was just at a meeting where they were telling me they had like 20 certified carpentries instructors or something at their library so it's really kind of exploded. So Tim has been really involved with the carpentries. So maybe just a few words about the carpentries movement and their pedagogy and a few thoughts on the importance that plays especially for those who aren't familiar with it. And maybe what sorts of things librarians have been doing in that program? Oh, yes. So as you mentioned, I think librarians have been involved in the carpentries for a good while and various roles but so the carpentries is a global, it's a volunteer network of instructors and we teach like basic and computing and data skills to typically researchers but now information professionals, librarians, research staff on how to do their research better, how to automate tasks, how to be more reproducible in their work. So all these are kind of wrapped up into the carpentries and there's three carpentries, the software carpentry which teaches best practice in software development techniques because a lot of researchers get into doing their research and they have no background in CS, they don't know how to write scripts and so it's teaching some of those best practices. Data carpentry has geared it. More domain based so there's data carpentry with genomics and social science and geospatial and it's geared to the methodologies for those specific domains and dealing with too much data that researchers don't know how to handle it. How to manage it, how to analyze, how to do data visualization and the library carpentry which I've been involved with on the governance committee and now we're the latest carpentries so there's three carpentries and then that's more geared at teaching librarians how to code and deal with data, how to kind of automate their own work processes and then be able to better engage and communicate with technical staff in their libraries and also be able to engage better with cutting edge research, right? So those are kind of the three carpentries and if you wanna become an instructor you go through a training program that is really not about technology it's about how people learn and how to teach and we use kind of education psychology best practices on those kind of facets and you go through a two day workshop and then you become certified as an instructor. I think that's, and then libraries, yeah so the library and all the library carpentry has been like three or four years old and we're, I think there's a lot of also members of, because carpentries is a membership organization a lot of libraries that are members of the carpentries so it's kind of taken off and like within the UC system I think we have like probably 20 librarians who are certified and we chat, teach workshops together so it's kind of an active environment. Yeah, one of the cool things about the library's collaboration with the CDS, the Center for Data Science at NYU is that we actually split our carpentries membership so it's supported by both organizations which has been great for me because I don't have to do all the administrative work I can share that too but also just this deep understanding that like it's a great place for learning to happen in both of these multidisciplinary spaces which has been very cool to see. I'm sorry I should, I forgot something but one of the big components of the carpentries is we really want to have and it's called inclusive pedagogy where we really have a warm welcoming environment where people can come in regardless of their background and build confidence and have a positive mindset so they can learn and grow, right? So we have a really serious, we're serious about our code of conduct and we do have a real like, it's been pretty successful in this mission. So now we have, we're translating lessons into multiple languages, we have an African task force has been really incredibly active and wonderful like the work that they've done in the continent of Africa. We've had workshops on all seven continents so we really have a global reach and it's really premised on being inclusive and to try to improve the representativeness of people in computing and data science as well. So yeah, so anyway. So before I open it up for audience questions I want to just go through the panel once and give you an opportunity to mention maybe just pick one sort of new and upcoming program service training offering that's code related that you are, that you have in the works for the near future. So that's relevant for you. Just to give you a chance to promote something, you know what it is you're up to. Matt, you want to start? I mean, do you spot? Your job's a little different but... Yeah, I don't have any, well, I mean, you know, I'm trying to, you know, I guess a new thing. One of the, actually, I'd say it's not a specific, it's an infrastructural thing that I'm working on and so it's not an educational offering but it's, one of the things I'm trying to do is deploy, you know, get a university-wide Jupyter Hub deployment, you know, and I love Jupyter and it's one of the things that brings people, you know, they're like, oh, I heard about this Jupyter thing and I need to learn it and then they learn it and it's, you know, they've, like, changed my life. I've had a couple people come and say that and I was like, oh, wow, that's a nice feeling. And, you know, so trying, you know, I've been working on kind of getting a large-scale deployment so that we can do teaching, you know, because as a platform for teaching, it's just, it's really fantastic. The computer scientists don't understand it or like it but everybody else, the scientists and the social scientists and everyone are like, well, it's just a no-brainer. So, you know, that's, and they're the ones who are like really using it so, you know, trying, you know, building and so I would say one of the things that I've been, the thing that I've been looking at and kind of using as inspiration is all of the work coming out of UC Berkeley, both on kind of the infrastructural Jupyter hub deployment stuff that they've done but also they are developing curriculum materials, you know, in and also talking about how to teach with Jupyter notebooks and so like all of that stuff. Everything that's coming out of Berkeley, out of the Berkeley Institute for Data Science on the stuff that, it's not something I'm doing but it's something I draw on so, yeah. That's the data aid course. Yeah. Yeah. I'm also, I want to address like one major challenge of supporting data science if I may, which is something I want to have a conversation about. The hardest part for me in supporting data science is collecting, like actually providing data. A lot of the data sets that students email me how like a Google site with a million images of a face and the person who runs it needs a signature from a department head but we might want to collect it and give people access to it in the library and it's one researcher maintaining that and they're not equipped to deal with like library's collection stuff or even like be willing to have a librarian broker that agreement. So for me, the thing that I really want to work on moving forward is almost less on the pedagogy side and more on sort of the maintenance and infrastructure for collection to support data science. So that's like my major challenge that I'm hoping to address. So we at the URI AI lab are interested in K through 12 programs, especially in machine learning and artificial intelligence. We have just hosted our first AI summer camp for kids ranging in age from seven to 17. It was a huge success and one of the summer camps was absolutely free and we identified there was a need for AI education among this group of students who belong to the underserved communities. We definitely identified a need there and we are working on a couple of projects. One of them is an after school program where the students would come to the URI library and have this two hour session instructed by one of us on artificial intelligence and machine learning. Apart from that, we also reach out to schools. We do talks and just raise awareness about artificial intelligence and try to break some of the myths about artificial intelligence and machine learning. Yeah, I was trying to think. One of the, so this class that I mentioned on mass incarceration, they're gonna be doing a second quarter which will they'll have their projects and we're gonna do some more teaching and I really enjoy teaching R and mapping which is I'm gonna formulate that into regular workshop because I like getting away from proprietary tools like Esri's, you know, Gripbone, Geospatial. And the other thing I'm talking with some colleagues is about like data ethics and ethical concerns around machine learning and these kind of black box systems of algorithms and we're talking about having a workshop series around that. Okay, so are there questions from the audience? I need to bring the mic out because we're being recorded, so great. Thank you again, that was an awesome presentation. Quick question on how you entice students. I know you said they sort of come to you but do you and you're talking about outreaches, are there ways that you reach the students in non, say the non-computer science, non-engineering students, some hackathons, those kinds of things and how do you physically do it? Do you actually go to the colleges, go to the classes, sit in a breezeway and wait for people to come to you? I mean, I laugh but I do hold office hours and I get a lot of people that way so standing in the breezeway is not a bad choice. I actually mainly see non-engineering and non-computer science students in my like intro machine learning and intro to getting things classes. A lot of times it's the students outside of those disciplines who feel like they're behind and wanna catch up even though it's not the case in reality. I'm sure the computer science students would benefit quite a bit from coming. So I actually think that when I'm looking at sort of my metrics of who's attending, it's mostly master students by far and mostly in non-engineering and computer science disciplines or people doing computational research but not in COMSI or like computer engineering. I have a few things to add to that. We do actively contact professors across campus and try to advocate data science and machine learning to be integrated into their course curriculum or at least talk to their students about this workshop that are available so that they can attend this workshops. And we advertise them heavily through our university newsletter on every digital media that's present across the campus. As far as our outreach programs are concerned, we work with Office of Diversity and by now we have several contacts with schools directly and they have approached us to coordinate with them and then we also attend events all across Rhode Island. I am just the dean of graduate students. I'll just, the deans, whenever I run off of the workshop, there's just a forum to sign up so I'll just email the various deans and say, hey, distribute this to the grad students and then it just fills up. And actually last term I sent it to a waiting list and it filled up after a day and I actually have no idea how they found it. I think the demand for the computing, like the Python computing stuff is so great that I just put it out there and it fills up which I actually don't know some of the under, I think it's just like every term it'll be a different, like it'll be engineering heavy or business heavy and I think it just, it gets circulated amongst the graduate students. But the way that I've started is by emailing like the dean of grad students or like the dean of research at a different school and they'll distribute it. Usually I ask them, can you distribute this to the grad students and that's how I've been reaching the graduate students at least who have been the primary students, people taking it. So let's take one more question because I think we're just about out. Thank you for an excellent, excellent panel. Jane Greenberg, I'm an high school professor too and we have, we've incorporated data science, we have data science certificate now and I lead a program called Leeds Library Education and Data Science. We've now graduated 21 fellows in the area so we're really into this space but it's a little self promotion but sorry, we'll spread the joy, right? But my question to you and anybody in this room is dealing, not everyone here is a librarian but we are seeing these services needed in libraries and how do you get this into things like the ALA accreditation? These conversations need to happen. We just got through our accreditation and it's a process and we all, for people who have an accredited ALA degree it's still very important. I know some schools don't do it anymore. There's a long debate there. But how do we, this is really important for future librarians and so I would be interested in any comments people have on that. Thank you so much. So we're doing that, we're going through accreditation right now and so we'll see how ALA responds to some of the changes that we've made to our curriculum. But I think this, for ALA the standards, there's the set of standards and you can describe how you meet those standards and what evidence you marshal to do that. So I think part of it is there's interpretation that happens in what those standards mean and I think we'll see how they respond to the changes that we've made. But I think we just have to, I mean I personally think there is some flexibility in that accreditation process. I don't think it's in ALA's best interest to not accredit previously accredited institution. I mean, so I think maybe there needs to be a little bit more, this is being recorded, but a little bit more risk taking on the part of the faculty who are doing that accreditation process. But it is, I mean the problem with that is the process is so incredibly, it's not, it's very involved to, there's some aspects of it to the credit, but there's some of it that it seems like it's just a lot of paperwork and so like maybe, I mean this has been a reoccurring conversation in the high schools about how to streamline that process and make it more flexible. But there's a lot of programs that are accredited by ALA that have this data science. I mean Michigan is a good example there. The whole program is accredited by ALA and you can go through that program and take only data science classes. So I think there is the space there, but it is a hard process. That's a little scary that there's only, they can go through to get a library degree with only data science classes. That is, it's ALA accredited. It's ALA accredited. They don't typically go to libraries. They go to Google, but the whole program, that's a structural configuration. God, next panel is gonna be about that. So I wanna pull that thread for like five, four hours. Let's take one more question. At Pitt, you have to take, at Pitt though, you have to take the core library. It is 10, so people wanna run to break, but we have. Reference. Good morning, my name is Teresa Byrd. I'm Dean of the University Library at the University of San Diego. I am on the board, the accreditation board for ALA for library schools. So I certainly enjoy hearing your comments this morning. Not sure you can, we only accredited the MLS programs. There is no MLS program in Michigan, it's MSI. Then there has to be a track or something because I'm agreeing with her. We wouldn't credit a program that is not solely not. But you don't credit tracks, you credit programs. But your comment from the library school faculty member, I think whatever you would like, I can, you can share your information with me. You can give it to Karen O'Brien. We are looking at the process, I have to tell you. I read all of those documents and it's a lot of stuff to read. So I am certainly in favor of streamlining the process and we are looking at that right now. So thank you, I just wanted to share that you had an accreditation person listening to your comments. Thank you. All right everyone, thanks for coming. Thank you. We're all around and feel free to chat with us. If you'd like.