 Hi, this is Andrew Lee. We are at the Wikimedia session on 10 years of wiki data How it all started and where it is going and we're here with Lydia Pinscher and Danny grander chicks Hi, I'm Lydia doing the product management for wiki data and have been with wiki data since it started about 10 years ago And I'm Danny. I am with the Wikimedia foundation working on abstract Wikipedia and wiki functions And I've been the founder of wiki data in those 10 years back Great and Danny since you're the founder of wiki data. Tell us a little bit about how you remember how it all started The very first thing that I remember is about this is that there was a call for wiki mania 2005 and it was in Frankfurt and Both me and Marcus Crutch back then were PhD students in Karlsruhe and we both were wikipedia. We knew that so we're like Oh, I want to go to the c conference. So we saw the call for papers and let's submit something where we connect our research topic Which was a semantic web together with wikipedia and we didn't exactly know what it means, but we'll figure it out So we wrote a paper submitted it to Frankfurt. It was accepted. It was actually the first presented at a conference and this is how the semantic media wiki idea started and We were from the beginning. We really wanted to make it a semantic wikipedia But this then took a few years to actually get to that place My first memory is we're sitting in Karlsruhe in a cafe with Benny To talk about this crazy project he was going to do and that I should be a part of it to help with Community support Helping wikipedia and understand what the project is about and so on and also making sure that the development team knows What's going on in the community? Yeah crazy times and really cool first job right out of university Lydia then was also working on semantic media wiki as Community manager and so I knew her from there and I really trusted so I was approaching her basically with this role of helping us with the community in wiki data and I'm super glad that she accepted Yeah, and so it's a magic media wiki definitely was part of the community in Providing kind of database like or structured data type functions to the traditional media wiki But how did wiki data get its footing in terms of Approval as a project within the movement and then why was it with wiki media Deutschland? Marcus and I were trying to push for Semantic wikipedia to happen for a while and the project has changed over time So originally the idea was actually to make each of the wikipedia's individually semantic to add Structure data to each of them. It was I think Eric Muller who later suggests that we should have one wiki data There the different wikipedia access the data from and he also actually suggested a name he He got a domain early on before even the whole idea happened. I think he had it also like since the mid 2000s They had in general like Agreement from the community that this would be a good idea, but no one really believed it would happen And it took us years to really get to the point of we had let's make it happen And when I finished my PhD, I was like in the point there. What do I do next and one of the funders It was the N Institute for artificial intelligence Approach me and said, you know, you can actually now have the opportunity to make this happen for real You're you're in the right time. You don't have any other obligations right now And they would provide some of the funding I was then I had to find the other funder so a squared gave us half of the money and I approached Google and they agreed to give us some of the money a quarter and the wikipedia foundation managed to get a Quarter secured from the Gordon and Betty Moore foundation and then we approached the board of trustees but it would be okay to do this project and The board of trustees agreed after talking with them and explaining what is what is is and then the question was where to host it Eric was very careful telling me that the foundation at that point in history Wouldn't be the right place to host it that he doesn't he didn't feel like they could pull off this kind of project and we were exploring a few different options and One of them was to have an organization of its own the money would have been sufficient to actually pull that up Another one was to approach the committee at Deutschland and ask them They would be interested in hosting that and power Richter who was the idea then Agreed for it. It was actually quite enthusiastic Understanding that this would be a good opportunity for the committee at Deutschland to Professionalize and change quite a lot by bringing in such a big number of people great And one of the reasons why wiki data had maybe its initial Success of the community was because it managed the site linking which was very manual process, right? Before wiki data you'd have to put a long list of links the bottom of a wiki data I'm sorry wikipedia article So tell us more about that that uptake by the community and how site links probably was the entrée for a lot of the community to wiki data so Sightlings as you say maintaining them in each wikipedia article across all the different language origins cause a huge amount of Education and work that wikipedia didn't like rightfully Starting with that as a way to solve that one problem for wikipedia editors as Especially was really helpful in in gaining that acceptance on one hand and and Supporting wikipedia editors especially but it also helped would strap wiki wiki data because By creating all these items for all the existing wikipedia content We suddenly had a lot of items on wiki data to start from with really important concepts to cover Yeah, we bought a lot of good will basically with that first step and we Changed a little part of wikipedia quite dramatically Some of the things that you could easily see for example in the charts was that the number of edits because of wiki data Dropped dramatically after wiki data has launched and the sightings was switched on which first led to a big scare And the people understood oh well is because the bot stopped editing the sightlings and keeping the raptor data across all the free-handled language editions And but this also means that we made Human edits much more visible and brought a motor to foreground and all these things and There it was in many smaller things wikipedia doesn't understand this travel But in many smaller wikipedia's the sightling edits Basically the majority of all edits and you looked at the history of an article you just saw bots editing the sightlings and you didn't even see who else was the editing and This is something that But and also they were inconsistent. They were basically always inconsistent All the sightings to each other. So this is where we bought a lot of good will for wiki data And where I think the first time people realized. Oh when this is not just some Crazy idea that won't happen anywhere. Anyway, it won't go anywhere But this is actually something that's really happening and it's impacting us right now Right and it's centralized things that made sense to be centralized, right? And then there are no statements at the beginning just the sightlings But then how did it grow from there and the early adopters who are the early adopters? so among the early adopters were definitely people who were Doing work on wikipedia that was now much more well-studied on wiki data so people who were working a lot on categorizing things on Putting in info boxes and updating them and so on As opposed to the people who care more about actually writing articles and researching them and that's all on so I think that was a bit of a Move of the type of people who really care about data as opposed to writing along from pros especially at the beginning Also, it was at wikimania in Washington where we asked like which of the communities would like to join us first so it was completely voluntarily in the beginning and The we got reached out by a number of wikipedia community is quite large one It was Hungarian Hebrew Italian Communities that are traditionally open to this kind of changes and they were among the first wikis to switch on Sightlings to allow access to wiki data and have it be there in production Which was really important for us because it completely validated what we were doing Right and today you see like Cadillac and Basque all these folks using it in ways that that really benefit those smaller communities Which is great and this is a big question now Like how is wiki data changed in the last 10 years since its launch and and then how does it change? Maybe the wiki media movement and wiki media projects Lydia, right? so I think one thing to understand is that at the beginning of wiki data is which really perceived as vaporware by a lot of people, right? it's this crazy idea that this team small team of people had and that they were working on but Maybe it's never going anywhere. Maybe it is we will see and it was not very well Understood what it's going to be how it's going to work What kind of impact this has it's going to have on wiki pia and the other wiki media projects or and even beyond the movement So a lot of people projected Their biggest hopes into wiki data and some other people projected their biggest fears into wiki data and one thing that changed as we develop more and more and as the community was building is a Deeper understanding of what wiki data actually is and where it is helpful where it is not and what it can be in the future Yeah, but we originally launched wiki data I basically didn't have almost any features at all You could just like create pages with labels and descriptions and sightings and the sightings are not even used anywhere So if there are no statements, there was no wiki data query service. There was no well-defined API and all of these things and Today we think of all these parts, you know as crucial components of wiki data, but they all launched later So wiki data has been developing over those 10 years thanks to Lydia's leadership In a really good way that led to more and more uses being able you know creating the wiki data query service Which personally I thought was completely impossible. I'm still amazed that this happened ever I was my original plans of wiki data didn't involve it because I didn't think it could work. I was wrong I was happily I was wrong And then editing lexicographical data to wiki data, which is a big thing And I hope this will be able to help the victories a lot the And obviously statements of all the different data types and the usage of descriptions in many many different places Not in the English Wikipedia unfortunately, but in many other places All of these things are I think helping wiki data grow and helping also the movement grow I Mean as we speak right now the the object count or the item count is 99.6 million in wiki data as you said to be able to query that and come back within seconds is pretty astonishing Even after we five plus years of experience with sparkle and wiki data query Which brings up the question like where do you see it going from here? Certainly questions about scalability and The stability of services going forward. Maybe Lydia you want to you're in the throes of this You're in the middle of it. So maybe you can answer that first. I could try So I think over the last 10 years wiki data has become More and more useful and more and more important for technology we use every day And I think that should and will continue right more and more people will in some shape or form Get knowledge through wiki data if they know it or not and What I would like to see is To get us to the point where more and more people also contribute to that knowledge maybe not everyone but I think we have a lot of opportunity for a lot more people to to share in in Building that knowledge base that we have and maintaining it So I think that is one important piece The other important piece is that as more and more people get exposed to our data in some shape or form in Applications and visualizations in services. I think it's really important that we Are aware that it's becoming much more important and much more Crucial that our data is is good And it's verified that it's complete that it is up to date and so on and at the same time the pressure to manipulate it just like on the big wiki peers is probably going to increase Because it's going to be a much more interesting target. Let's say So I think that is really important and the other important piece I think is that I don't want wiki data to be this one place where open data happens, right? I want a large connected network of places where linked open data happen and that work together and that are interconnected and that share data and That is all the work we are doing with the wiki-based ecosystem where we don't just have a wiki-based installation for wiki data To run this one big instance, but have other people other institutions other organizations be that museums or libraries for example Coal open culture projects that all run their own wiki-based and then interconnect that with wiki data so that wiki data Isn't that one place where we're all open data has to happen because I think that's not a good position to be in Great, thank you. Yeah, and I think something most people may not have seen if they weren't following it is the Making of wiki-based its own product and ecosystem and community and cloud service, which is pretty amazing as well Yeah, fully agree with Lydia and everything she said in addition I think something like the wiki data query service Needs to figure out how to keep the role it has been using as for growing wiki data and We need to figure out How to provide the services that the wiki data query service is currently providing also is stable in the future Given that we're probably going to grow even further What I am really excited about is how the dictionaries can use Wiki data and grow from there and there's so much duplication across the dictionaries and I think this is something that can very much happen in the next year or two and make a big exciting change there and And finally also the wiki site project is something that Excites me a lot and we really need to put resources into that one. There's so much interest from the community, but Not so much uptake from some places that could help with making that really happen and really Grow into a project and obviously then there is the work on aspect Wikipedia and wiki functions Which completely relies on wiki data, but we will have our own session during wiki mania on that topic Great. Thank you. And what were some of the best memories of working with wiki data over the years I think Denny the the story you told of why we have Q numbers and wiki data I'll leave it as exercise for the reader to look that up, which is a great personal story But Lydia what is what are some of your best memories about wiki data? right, so as we were building after project and as we were Supporting the community and finding its own way one of the things I always did was When something happened in the world something important that grabbed people's attention was How long would it take for someone to actually update that particular data point? in wiki data and can I be faster than anyone else and there was this point with the election of Pope Benedict where It happened the first time where it was not fast enough and where someone was much faster than me And I was kind of the point where I thought like yes We made it this this is gonna work. This is this is gonna be great I Fully agree with Lydia again. I'm just those stories that are things that have nice My favorite example was when I was flying in Europe. I forgot the airline But I was sitting there. I was taking a look at the at the entertainment system with the inflight information And suddenly it was saying oh, this is powered by wiki data, and I was like great I'm in an airplane, which is powered by wiki data I'm not scary at all, but it's also like super amazing that those things happen. It's like wow Yeah, the the world out there is using you know, not just the Googles and the series of them and the apples of the world, but but also Well, okay an airline isn't particularly a small company but you know also small organizations are using this knowledge and doing all kind of fun things with it the other thing that they're really wiki data is the community and It has teached it has taught me again and again To be humble and how much better the community is than any single person really It's I Remember distinctly when we were looking for funding for wiki data I had a long discussion with a potential funder who was completely aghast by the idea that we would let the community decide on the ontology decide on Properties and classes will be available to wiki data because it's just will never work you have to pre-define that you can't have that done by a laborers and so on and We eventually lost that funding because I think now we have to leave that with the community and Leave it there Because we wouldn't Budge on that. So this is not not something that I found was negotiable at that point It does keep impressing me So so one of the things my favorite added to wiki data is For the dragon spike from the my little pony show And there's an edit that changes spike from being Dragon to being a fictional dragon and this just you know It shows how much the community cares. It shows the the quality of the ontology. It shows So many things in this little edit. It's just amazing besides being also genuinely funny I think Yeah, and I think that is one of the really core strengths of wiki data right that it is so Flexible and open and that a lot of people can do things with it that neither then, you know I had ever even dreamed about and that is super powerful and and always impressive when someone does something with it and you're like, oh Okay, I've never thought of this, but it's really cool Well, thank you both of you for your stewardship of the project that the amazing success that's out there So we appreciate it very much and we hope folks will attend abstract Wikipedia and other talks and be immersed in wiki data during wiki mania All right. Thanks everyone