 Okay. Good morning, everyone. I hope you're well. It's very exciting to be here, actually. And well, I almost didn't make it. So yay for that. And I'm especially excited to present today because this is kind of a personal journey that I took. This is the results of my PhD that hasn't been published yet. It's been submitted and it's really at the final stages, but you're the first to hear the results of it. So I hope it's interesting and feel free to ask questions. There's going to be a section at the end. Essentially, I just wanted to give you some intro into how it came to be and why I pursued it and then just describe some of the findings, not all of them. It's a 200 pages thingy. So it's kind of hard to shrink, but hopefully it's still going to be some highlights that you may find interesting, some things that surprised me personally. And let's just delve into it. So the first part is just giving a bit of an intro. And I said it's a personal presentation. And in a sense, this was sort of like giving birth. Essentially, I sent this proposal for the PhD to the Tel Aviv University in April of 2018, five and a half-ish years ago. It was from hospital. And a day after, I delivered my baby library. So this journey has been about growing the PhD and growing her and myself. And this is, by the way, this is a Lyra. This is a instrument. It's also a constellation. So that's the meaning of her name. And just to give a context to that, I'm an educator. I specialize in technology and learning. And I've been basically deploying or implementing Wikipedia and Wikidata in my academic courses at Tel Aviv University since 2013. I've been experimenting a lot. And besides my courses, I've been also assisting a bunch of other people to implement Wiki projects into the academic and educational curriculum. And when I started to research it, I found out that there's tons of, you know, information out there, academic research, basically discussing what Wikipedia can do as a learning platform, all the benefits that it has. And there are many in the literature. There's also quite a lot of academic literature back then, right, on Wikidata. But most of it was technical. Most of it was about maybe the ontological side of it. Less so about the potential that it holds for education in the wider sense of it, right? So not just in the classroom, but really as a lifelong learning platform. And that is what interests me. And so I submitted a proposal at the beginning of the academic year of 2017-2018. And in April, I submitted it right from hospital. It was approved in August. And so the next year in September, I started this journey. And this wasn't just my journey. It was a journey with this community, actually, without whom this research would not be possible. And when I started out, it was one of the first things that I did after the literature review and all of that and looking at the lit, is create this conceptual framework. This is at the beginning of the way, back in 2017, is how I imagined this platform operates. So from where I was looking, I said, okay, we have this technological platform, amazing. We interact with it in various ways. Back then, I thought, oh, there are probably two main interactions, right? Like data curation when we add information to the platform, just like with Wikipedia. And there's also data extraction when we try to query wiki data or extract information from it. So at the beginning of the road, these were the two main things that I saw. And I said, well, maybe probably there are going to be other interactions that we do with the platform, but these are the two main ones. And from these interactions come out uses. I mean, people use the platform in various ways. And I was assuming that from these interactions, learning happens. Also looking at the literature, I said, it looks like there are four different things, aspects that influence this learning or these interactions. And I map motivations and the skills that users have as external, as internal aspects that influence this interaction. And I map the community and the technology itself as external things that may influence this interaction. So this was my conceptual framework. This is how I thought it worked at the beginning back then in 2017. And essentially, interactions lead to uses that lead to learning. That was the sequence that I thought things were happening at. And that kind of led to four main research topics. I mean, under each here, under each topic that you see here, there are multiple questions in the PhD. I didn't want to put too much into it, but just to give you an essence. The first one, it really correlates to the conceptual framework. So the first one was just saying, okay, there are uses out there. We have no literature about it. Let's just map for the, as a first step, what is out there? Like what are the different uses? What are the projects that people have been doing around the world? And let's extract uses from them. And also let's look at the benefits and challenges that these early adapters of the platforms have been finding. And that was the first research questions. The second was about the interactions themselves. So let's map these interactions. Are there really just two or maybe there are others that we need to kind of know about to understand it better? So the second was about the user interactions and how the user is actually engaging with the platform, perceiving their learning potential. The third one was about these four aspects that I thought was influencing the interactions, the motivations and skills, the internal ones and the technology, the technological platform itself and the community around it as external influences. That was the third thing that I wanted to look at. And then the last research question was about having a holistic view, right? Looking at this whole conceptual framework that I showcased here and checking statistically if, am I right? Are my assumptions correct in looking at the different connections between all these components that I knew existed for sure, because I was part of the community. I knew that people are doing amazing projects. I knew that they're interacting with the platform. I knew that learning was happening, but I wasn't sure about these relationships. So these were the four research questions and I took in terms of the methodology. I'm not going to delve into it because it's quite boring, but I do have to say that it was a mixed method. It was mainly qualitative in essence, but I tried to have as much as quantitative data as I could. And essentially I had three phases for collecting data and then analyzing them. So I started out with an exploratory, as we usually do in long-term research projects like this one, and I interviewed seven well-known people from this community and the projects that they were doing. And that was the first phase and really analyzed to hell everything that they said in these interviews. Then I devised a questionnaire and 120 people, 121 actually from the global community fielded out, which was amazing. I didn't expect it originally. I thought I'll have a bit less. And then out of that, almost 16 interviews, which is usually, it's a high number for a PhD, usually five, 10 interviews if you really want to, but I felt like I need to do more. And I ended up interviewing half of the people who actually participated. And you can see if that's interesting to you that the questionnaire was quite long. It was 71 questions dealing with different aspects that I mentioned before, the interactions, the motivation skills, technology, and community. And also some personal data, because I was interested to learn if I can learn something about where people come from, if they have specific skills, background that is helpful, et cetera. So I was interested in that socio-demographic aspect of it as well. And just to give you a highlight, plus minus I had 67% male and 30% women, which is better than the rate that we usually have for editors in Wikipedia. So I was happy about that. I was happy I was able to recruit women. And you can see a small map below showcasing where these people came from around the world. So I really made an effort to be inclusive on various levels and just include people from different places in the world. And this is, in essence, dangerous. So let's talk a bit about what I found. It's a lot. And I'm going to say disclaimer, because you're Wikimedians and I know some of you are really interested in the topic and I've received questions in advance. And people wanting to know and learn these slides include much more details than what I'm going to talk about right now. And that's for you to be able to delve in later on into some more details if that interests you. So I'm going to go very quick over the three first research questions and delve into the last one of the holistic views of whether I was right about the connections, et cetera, and some of the findings that I saw that are interesting. So in essence, with the first question of just mapping up the uses, the benefits and challenges, we did that. I found essentially that there are around 10 different uses for Wikidata. And most of it is actually published already in an academic article that was published earlier this year. So if you want to delve into it, it's actually online. And if you're unable to access it, just ping me and I'll send you the PDF. I wish I was able to publish it open access. That's in parentheses, but it costs really a lot of money. So I wasn't able to do that at this point in my academic career. In terms of the interactions themselves, I can say really on a very high level that many, many people that are interacting with the platform find it very useful in terms of learning. I've mapped specifically what happens in data curation, data extraction processes, and actually compared what happens if you do interact with the platform in data curation or not, and the results were quite significant if you interact in either in data curation or data extraction, learning will happen much more than if you don't. And then part of that question was just mapping additional insights that people had from the interviews, right? Because the interviews were semi-structured, unlike the questionnaire that had very specific questions. So I was able to tap into people's minds and what interests them. And there were four main topics that came out that were not part of what I mapped originally in the questionnaire. And the first was about the really complex workflows. And again, I'm not going to delve into them. But if that is interesting to you, save it for a question in that is coming up. The second theme was introduction to newcomers that is still very complex and need to be addressed. Then there was an aspect, a theme that kept recurring about using Wikidata in the classroom as a learning platform in the context of education or academia. And many people found that it's really helpful in terms of developing overall skills that people need specific knowledge as well as data literacy, which is really interesting. And there were also, there's also, because these are Wikimedians, and I also asked, there was another theme, the last one that I mapped, that it was outside the scope of the questionnaire was suggested changes to the platform. This was basically, if you had the key to the realm, the keys to the realm, what would you change in the platform and why? And many people had lots of insights about the UX, about tools, about various really interesting things. I want to jump to, I then mapped, of course, the four aspects that I mentioned. So the motivations, the skills, the community and technology with each of them, there were multiple mapping, some of it came from the questionnaires, some of it was mapped from interviews. But I'm going to skip it. And if you want to delve into it, you can look at it after. And I want to move to the fourth question of looking at an overview at the conceptual framework that I had. And what I found is that there are specific correlations between different components that I didn't really expect. Funny enough, or, well, it's not really surprising. It's also in the literature, actually, but two motivations seem to be a really strong component. It's connecting to both learning processes in the community itself. So these were the two strongest correlations that I found. Interestingly enough, the skills people had before they joined were not necessarily correlated, which I think is excellent news because that means it doesn't matter what the skill set you have where you enter the room. Yes, it can be helpful, but it is not necessarily indicative of whether or not you'll learn. So in essence, whatever the skills you have, whatever background you have, if you continue at it and you persist, you will learn new things through these platforms and you will improve your knowledge and skills. And I want to jump to some of the predictions. So again, there was an assumption that interactions lead to uses that lead to learning and that there are these four aspects that influence it. And I did something that is called directional analysis. And I just wanted to see if indeed these four influencing aspects could really predict learning. What do you say? What do you think that I found just from the instinct of you using Wikidata? Any thoughts in the room? Yeah, interesting. Okay. So what I found is that motivations, technology and community can positively predict learning with motivation being the most influential. I thought that was interesting. And again, the role of skills is not critical for predicting learning. That is what we saw from the prediction analysis. And I think I've talked about that. And I'm guessing this is the really interesting part, the mediation analysis that we did, a path analysis to kind of see if the model really fits. So the model fits in essence, but the connections are completely not what I figured out at the beginning. And I guess what we figured out is that there are interrelations that we didn't expect at the beginning, that I didn't see like I saw at the beginning like a very clear sequence of events. And it turns out that some of the influencing factors are directly influencing learning, which I think is interesting. And one of the key things out of this was that the community came out as a really strong component in predicting learning. And by far, it means that the more that you're connected to the community, the more likely you'll learn stuff. So the community, again, is a really strong component in this. And I guess it would be easier to see visually, right? So this was the beginning of the journey. This is how I thought the sequence works. But really, it's more like this, right? Like there are specific things that influence directly. You can see the green is the correlations. The purples are predictions, like what can predict what. And you can see some direct links that do not go through interactions. It's just directly motivation, directly influenced learning, etc. And there is also some reverse prediction that again, I didn't expect. So it completely changed the way that I saw the model at the beginning, which was, I thought, a nice surprise. And I want to share a few things that I found out in this or talked about in the discussion after I mapped out all the different findings. And first of all, again, it seems like there are interrelations. It's a complex picture. It's not necessarily in a specific sequence, as we expected. And everything kind of influences everything else. It's like an ecosystem. And it seems like data visualization, the possibility for wiki data to help us visualize data is a critical component to allow high order thinking, critical thinking, and developing various skills. It also mapped various challenges with the platform. So it's not that I'm saying this platform is perfect. Absolutely not. There have been a lot of discussion in the PhD about the challenges and the issues that we have. But I would say very clearly it comes out that it does help to improve both digital and data literacy to improve it. And in essence, it is an important component in open education that is sort of an untapped potential. And we need many more people to actually engage with it than we have so far. In the conclusion, I actually mapped the value of wiki data for the world. And I think I'm going to skip it because for you wikimidians or wikidatans, this would make perfect sense. And I will say that the only last thing that I talk about in that section is generative AI. You've heard me talk about that before. If you've came to the gen AI panel, but essentially wiki data as of now is not feeding all the big LLMs. So there is a whole component that needs to be addressed of connecting linked open data in general and wiki data specifically to the realm of large language models and the whole idea of generative AI. And we need to explore further how we can use these generative AI tools in order to enhance our work. Some of it we're already seeing, or at least I'm already seeing in my classroom, for instance by telling my students, one of the things that I mapped before was that learning sparkle is a huge threshold for people. It's very difficult for them to learn, to understand, and it's preventing people from actually being able to use the platform properly. And now we can simply go to chat GPT and say, right, Mary, sparkle query that does ABCD. And it will produce most of the times a really decent code. And then if you know a bit, you can tweak it, you can change it, you can think about it critically, but you don't really need to be proficient in sparkle the way it used to be, or need specifically to be connected to the right people in the community who know how to do it. So this was that. And there's a bunch of suggestions at the end of my research for future research that need to happen from implementing, from, first of all, rechecking the model, the new model that I came up with, to actually implementing wiki data on a wider range around the world, specifically addressing challenges that we've seen, further exploring wiki base, further exploring reliability and bias in wiki data, and really harnessing new technologies to what we do. So all these needs to still happen. And I want to conclude with this hope. So this is what I hope this research will do or help do. I hope it helps inspire more people to engage with the platform to use it, not only the classroom, but really as a lifelong learning platform. And that we will continue to do this work. And I also had a personal hope because remember this was a really personal journey for me. I hope that I keep growing and I mentioned Lyra. This was the beginning of the process. And this is her just now in April, she turned five. So I hope we keep growing together and with the community. And that is it. I want to move to the Q&A and open the floor for questions. Not everyone at once. I know you kind of grasped, oh my gosh, I grew a human throughout the process through COVID. This is Laila Wikimedia Foundation research team. Thank you so much for your presentation. I have a question about the motivation types that you considered. Can you break it down and speak to what kind of motivations did you consider in your research? Sure. So it was a mix of specific questions in the questionnaire that we're dealing with motivations and trying to see the perception of users of what type of, to what extent do you think that this motivates you, etc. So there were questions in the questionnaire dealing with just people's perception of their own motivation, what motivates them, etc. And then in the interviews, I was able to delve in a bit more. And all the statements that I mapped, because it was, you know how it is for qualitative, you basically do an interview and then you dissect all the sentences and you try to map them, etc. So I mapped three subcategories first for initial motivation to participate. So there's a bunch of, and then subcategories of what that looks like, and I can share it with you later if that's interesting, then what motivates people to stay once they've went through the threshold. And the third category was what inhibits motivation, like what are the things that really hinder people from wanting to be there. And so each of them is mapped with some statistics behind it. And if you want to delve into it, you can definitely see more. I think I did put something, I might have put something here. No, I didn't. Okay. I think it's hidden probably and I can see it, but I can send it to you. Any other questions about the process? Silence in the room. Oh yeah, go for it. Not really a question about the process, more a question about like, I don't know, the takeaways. So I think that that I like the idea of thinking about projects like WikiData as learning processes, but it seems like there's like, I guess the question is, do you see there as being any conflict between like how you might design a process for like maximizing learning and the goals of, I don't know, providing really high quality data all the time, because I think that people are in the process of learning. If you're really maximizing for learning, doesn't that also involve making mistakes? And how do you, this is a very conceptual like question, but you've obviously been thinking about this a lot. How do you balance those two goals, right? The goal to promote sort of individual growth, perhaps through experimentation and mistake making mistakes and the goals of promoting high quality data. Yeah, I think that's a really valid concern. And really, there's no hiding from it. Like it's, it's not something that we can skip over. Part of learning and actually some would claim in the literature and beyond just common sense, right? And experience in life. The mistakes we make sometimes are the best or the moments where we learn most. So I wouldn't necessarily avoid mistakes. Mistakes are part of any project and anyone who's designing a project should include portions to deal with mistakes and to deal with what can go wrong. And you know, so designing for a, I would say, optimistic process would include time for review, time for evaluation and iteration on whatever it is that you're doing. Not including it and assuming everything is going to go smoothly is a recipe for failure, in essence, in my mind. So I think from what I've seen, the people who design indoor process of implementing Wikidata as a learning platform, some time for iterations, for questions, for feedback, for interaction is most successful. And I think one of the key findings here, the strength of the community in this case, in this specific case of how and think about it in the context of Wikipedia that has the exact opposite reputation, right? Of being sometimes notorious for not being welcoming, et cetera. And it seems like for Wikidata, it's the opposite. It's the community is a huge, huge component of succeeding in the platform. So the more you're connected to the, to other people who've been doing it, learning from how they model stuff, modeling has been a huge issue throughout the, I didn't have a chance to, to get into it today, but it really stood out as one of the things that people grapple with, as well as lack of tools, documentation, you know, all of that. I mean, these are things that any Wikimedia and any Wikidata would know. But, but really it's about just making sure that you are tapping to, to the community as an actual resource that can help you throughout your learning process. And it seems like the community is absolutely enhancing it. And many, many people, apropos motivations and Leila's question, one of the main components that I thought was really interesting in terms of initial motivation. And actually I can flag the stuff here, just because he's here, I think a huge portion of what I saw in the research came from his sort of campaign to teach the global community about Wikidata. And many marked, you know, coming to a conference or doing a workshop as, as a really key motivator to, to start. So community is huge, designed for, for mistakes and, and learning is actually happening in, in various levels. It was quite amazing to see. Last questions, because I think we need to go on a break. Yeah, we'll take, we'll take two more, two very quickly, then we'll, we'll move to a break, but we do have Mako also, you are part of the two of the prerecorded videos this morning. We're just, yeah. So we can also take maybe a couple of questions for Mako during the break if people want to stick around. Hi, Jonathan Fran from Wikimedia Deutschland. We're, be very interested in the onboarding new people and understanding what keeps them. Thank you very much for that. We mentioned the generative AI and we talked about it yesterday. I was wondering if you had known about any projects that were trying to use it and how they were unable to succeed. You mean trying to use generative AI? Let's just say machine learning in general. So there needs to be a disclaimer, right? Because the time that I actually collected data was way, way before this happened, this whole generative AI. And it kind of came about when I was in the final stages of data analysis and actually writing the dissertation. It's still included in the dissertation because, I mean, I'm highly interested in that and I see all these connections. So I was able to include at least highlights for it and trying to direct future research about it. But to your question, if they're existing, so I kind of differentiate between using AI, like old school AI, if I can say that, to generative AI. These are, to me, these are two different things. And in terms of AI, in essence, there have been many. I mean, the Wikidata community has been using various tools, some of them relying on AI to kind of do its work, to fight vandalism, to track various things. There have been efforts. I don't know if Andrew Lee is here, but Andrew, for instance, at the Met, one of the first, oh, I see Richard Ferris here. Hi. One of the first projects that I mapped at the original stages was the Met. And they have been doing some innovative stuff that I recommend all of you to follow. One of them was to create, for instance, the PICTS game. So that included, actually, machine learning and taking a bunch of images. There was a huge set of data, both the metadata and the pictures themselves that they worked in collaboration. I think it was IBM, Microsoft, or MIT. There were a bunch of people coming together to do this, and they trained what happens with these big museums is sometimes they have a really small portion of their collection really well mapped. And the rest of it lacks a lot of details. So they said, let's train the machines on the data set that we have that is really, really high quality and good. And we know that the data is, we have complete metadata on that. And they did that. And then they used what that the machine learned to, for the Wikidata game. So the PICTS game shows you a picture. And all you need to do is say, hey, is there a tree? Isn't there a tree? Is there an instrument or not? Is there a horse or not? Right? Like this micro contribution. So I think this is an untapped, I mean, all of the Wikidatons in the room who've tried it got hooked. Like it's highly addictive. And it's really easy for engagement. So I think that's an untapped potential for engagement with wider audiences. But that's an example of using AI to train models to then suggest things for the general public to kind of help us map the Wikidata. And the cool thing about the, if that's the first time you're hearing about the Wikidata game, it's not the only one. But what happens is that if I click yes, there's a tree in the picture, there's going to be behind the scenes and editing Wikidata that is adding a statement that this picture depicts a tree, which is quite awesome. And you don't have to know anything about Sparkle or Wikidata, you just play a game. And this is allowing engagement from young kids to adults to professionals to anyone on this on the spectrum. And I think it's one of the avenues we really have to kind of further explore the whole idea of gamification and micro contributions, like small contributions from the public to help us do the work of mapping the world. Last question. Hi, I'm Jack and I'm a Wikipedia and a teacher and maybe it's of topic, but I was wondering if you have some insights about how to use the Wikidata in education in schools. Oh, yeah, tons. I mean, this is, this is essentially what sparked where are you from? Can I ask Italy? Amazing. So I think you can find quite a lot of insights in the article that was already published because it maps in essence some projects that already exist and you can kind of get inspiration from and see the types of uses of how we can use it for education, but tons. Like I've, I've, I essentially opened in 2018, the first global academic course to feature Wikidata. So if you want, we can talk about that more. But, but yeah, the whole idea is to use it as a platform, especially in a time where we see definite changes in the way that the younger generations are consuming knowledge and also producing knowledge. So we have to address it as educators. We can't continue to teaching like old school ways. And I think Wikidata is very helpful in doing that. And before we had the Gen AI revolution happening, Wikidata has been actually used by Syria and Alexa. So people were using it without knowing and growingly in essence, where when you requested things from Syria, Alexa, it would more likely bring the result from Wikidata than from Wikipedia. So there's a whole generation out there that has actually been using it without knowing. And now, now it changes with Gen AI and we need to address it like globally. But, but yeah, it's, it's just a tool to enhance data literacy and make sure that people understand the meaning of data and engage with it in ways that can enhance learning and learning a context. There are amazing, um, this article maps different uses that I think you'll find very compelling, including learning in context, like the ability, for instance, to create timelines that allow you to see at the Prado, you can come from this Spanish speaking world sort of. So the Prado Museum, and some of you may know Hystropedia, but the Hystropedia guys actually help the Prado Museum develop multiple timelines for their museum. So you can actually really learn from context in a context. So you can see the artifact from the museum, but you can choose what happened in philosophy in that period and what happened in, and suddenly you can see a period and you can, you can see what happened. You can see something in context and you can unravel connections that you didn't know were there because it's linked data. So all sorts of things like that that really changed interactions between human and knowledge from where I stand and are essential for anyone doing work in the 21st century. And with that, I want to thank you all for listening and joining my journey. We'll take a break and I hope you have a good rest of the day. Thank you. If you have questions for Mako over the next five minutes of the break, come hang out over here and we'll get started again in a little bit. Thank you.