 Hello everyone. Welcome to Wikimedia Hackathon 2021 showcase. During this showcase, some wonderful people will be presenting their works which they did during the hackathon, so get ready to be inspired with some great projects which are done in really short time. Before we start, I have a quick announcement that right after this session to be more precise at 1530 UTC, we will have a jitsi meeting for social gathering, discussions, Q&A, and rest of the showcase for people who doesn't prefer to be recorded. So if you hang on to your seats, the first presentation is from Isaac and will be about visualizing Wikipedia link gender statistics. Sounds very interesting. Mike is yours, Isaac. Thank you. Hi all. I'm Isaac Johnson. I'm a research scientist with the Wikimedia Foundation and today I'll be showcasing the Wikigender script which is a user script that visualizes the distribution of men, women, and people with non-binary gender identities that are linked from Wikipedia articles. The script aims to help uncover systemic biases associated with gender and links on Wikipedia. I want to give some acknowledgments up front. The script was developed by user Taylor Robinson during this hackathon. The backend API that provides the data was developed by me, user Isaac WMF, and the design for the script that you're about to see was inspired by the human and key project. So as an example of what the script does, let's look at the English Wikipedia article for modern art. You'll notice the script inserts a button with the gender equality symbol on the right side of the page. Clicking on that symbol sends a request to the API that I built for the gender statistics associated with the current version of this article. And this is data that's all taken directly from Wikidata, but the backend API contains a snapshot of the data that is pre-formatted to keep the tool quick. All right. So when the data sent back, the script does two things. It adds the summary view at the top where you can see that only 11% of the links to people in this article are for women and there's no transgender or non-binary representation in this article. And as we scroll through this article, we can see that this representation is even more imbalanced in terms of link position. All of the links in the first half of the article are to men and it's not until we get to this kind of see also type section that we begin to see links to women artists. Now if I scroll back to the top and we switch to a different language, say Italian Wikipedia, the script still works. It's in though in its current form it has its layout fully in English still, but we can request the statistics look again. And here we'll see that the Italian version is even more skewed towards male artists. And again that kind of positionality bias exists perhaps even more strongly. So with that kind of in conclusion, I'd like to remind folks that this is a work in progress, but please let us know if you have any feedback on the design. But thanks for watching and I hope that this tool inspires some editing and highlights the importance of work done by groups like Women in Red to combat the gender bias on Wikipedia. Thank you. Thank you Isaac. I am a Women in Red volunteer, so personally I am very interested with this project. And our next presentation is from Luca Mauri and will be about Datatrack, which is a specific Wikipedia instance. Have your stay Luca. Hello, good afternoon. Thank you for having me. So this is very quick presentation for the moment. It's for a small project we are working for some time now. We are from Italy and we are a small group of trackers as you might have imagined. And we run the most complete guide to start track in Italian language. This is a project that is running since several years in different forms. But just recently we migrated a few years ago the project to a standard wiki that is called wiki track. And again more recently in the end of last year we set up a wiki-based instance in order to get the data for our wiki directly from a structured data source. There is actually Datatrack. Datatrack is the site you can see now. So it's a standard media wiki instance with the wiki-based installation. And it provides open data access to start track specifics. So it comprises of course data that might be present also in wiki data for instance. But since this is a specific one we can put in this site also data that are not acceptable, not broad enough for wiki data, but that can be more specifically put into a site like this. I can quickly show you around as you see it is a pretty standard installation. I can show you the setup of a very recent episode that went online. As you can see we are taking advantage of the multi-language capability of wiki-based. So we have of course data in Italian that are useful for our own site. But we will also keep translation for at least in some part of the site in English for instance and in German sometimes. So as you can see we have an organized bunch of data regarding episode for instance in regards to technical data on the episodes, the stars and all the navigation data, external links and also some site links organized in order to simplify the inter wiki links. We can have another look at for instance a fictional vehicle that of course this is a hero ship from one of the main series. As you can see we have a lot of data from it and a very complete site of site links in order to simplify inter wiki. This project is of course open to anybody. So we wanted to let everybody know that it exists. It can be consumed also for of course it is an open wiki-based installation. So it can be consumed from other sources and from other products and we of course welcome the support and the contribution from other users regardless the language you can contribute in. Data track is available on this URL you see here. We also maintain our statistics in wiki apiary. We have an entry in the wiki-based registry we have a twitter account and finally we maintain a github repository with some interesting custom-made lua code that we are using to get information from this site and to present it to our main wiki track. So this is for all the trackers out there to gather and came here to to help us or to take data from from this. So well thank you very much this is this is the site next your time. Thank you very much Luca. Our next presentation is coming from Lucas Wackmeister and will be about M3 API. I hope I read it correctly which is a minimal modern media wiki api client. Mike is yours Lucas. Thanks. Hello I have to admit that my hackathon project I didn't start from scratch at this hackathon but I took some code that I had lying around on my heart disk for a while and polished it and cleaned it up and split it up and added tests and then I called the result M3 API which is the minimal modern media wiki api client. So let's start with the media wiki api client it's a library that you use for JavaScript to interact with the media wiki api so you would create a session for English Wikipedia let me zoom in just a little bit ignore that and then you can make API requests action query, metasite info if there's an array in the value it converts it to a list automatically standard stuff and then you do with those API results whatever you want and what the minimal part in the name means is that this is most of what there is to it it's not a big framework like Pi wiki bot for instance or some other libraries which put a lot of abstractions above the API and then you have something like a function to get a page and then you can make edits to a page and then at the end there's a save method which will do something with the API that's fine there's nothing wrong with that but I just prefer libraries that are a lot closer to the API and so you use these actions directly and you know exactly how you use the API and so if you prefer that style then this is more the library for you so in that respect it's pretty close to the mw api in python but there wasn't anything similar in javascript as far as I could tell so now we have this and it's also modern which mainly just means it uses promises so you use a weight here instead of some callback thing a bit more modern is this which is an async iterator so the module can do query continuation for you and automatically follow the multiple results and then you can as incarnately iterate over that and at the end if you break out of the loop in this example then it stops making more API requests so yeah that's the name explained and so far this is just the version 0.1 I still have ideas for some features that I want to add but haven't added yet but already works is that you can use this from node.js and also in the browser you'll just have to if you use it in the browser you'll have to add origin star because otherwise you can't make cross origin requests and I'm hopefully going to publish this to npm.js later this evening but because I did a stupid and don't know how to use npm I couldn't publish it before so you'll just have to wait a few more hours for that but then if you want you can use this to build your own projects using the media wiki API in javascript and let me know if there's anything missing and maybe I might add that for version 0.2 or eventually a version 1.0 as well but so far this is just the beginning of this I guess thank you thank you Lucas our next speaker is Haas Medine Turkey and we'll be talking about developing a bot to add mess information from APIs to wiki data which is open site patience bot Mike is your Haas Medine I think there is a technical issue um yeah if the next speaker is ready maybe we can continue I'm just waiting a kick from the backstage okay then okay then we can continue with with Galder Gonzanes and the session it will be about Olympic data tone and this is a very hot topic so sounds very interesting have your say Galder hello thank you Nestlehan so some weeks ago I discovered that most of the information we have I'm going to serve my screen I'm going to serve my screen some of the information we have here like for example who won every gold medal or how many gold silver and bronze medals were won by any given country at any given olympics weren't at wiki data so I decided to start an olympic data tone and edit a tone so some people has been here some work has been done not every work and most of the things are not already done and some of the queries we are asking here are not even remotely impossible but one thing I made goes to upload every medal won by every team in the history of all the olympics like this is for example for the current yet finished olympics like gold silver and bronze how how much of them and I have uploaded this for every olympics summer or winter in all the history for all the teams and also some other users like Mr. Synergy has added all the olympia event to all the sporting events but we still miss some information like for example participation of every athlete in which olympic because this is quite messy currently and I tried with open refine but it was really hard because there are a lot of people and it's really hard for open refight to find them but well this information is here so I created a template in basque wikipedia that automatically takes this information from wiki data and so here is the basque wikipedia article about the this one the the united states of the 2020 summer olympics so with only this piece of code we can get this one and this is available for all wikipedia now but we still are missing here the information about who won what because this will need some information here some other things to do and some really complex queries that someone who knows about it should write so Listeria can can do this list automatically and so this is the the work done and let's hear the next speaker thanks a lot gather and let us try again with house a medina I hope it will work this time again and Mike is yours has a medina with our open citation spot do you hear me yeah okay so uh hello everyone so I will be showing some insights about what we have I have done during the wiki uh the wiki uh mania hackathon this year so to introduce ourselves we are a research unit from Tunisia working on data engineering and semantics in general and we are specifically focusing on wikimedia projects so this wikimedia research unit has been created through the collaboration between i3 e tunisia section and wikimedia tunisia as well as the technical aspects so as a research unit we have been concerned about the lack of citations in wik in the scholarly publications that are uh that are indexed by wiki data no digraph and so we had to search for a note for a semantic database that includes citation data and this database should be cc0 so that we cannot have some problems with uh related to uh legal barriers for the reuse of data so we have thankfully we have uh found open citations which is uh an open database for citations and it allows reuse thanks to its cc0 license so thankfully it has a restful api so we will use it so I have done this kit library repository I will share with you the code so I use it request to get the data from that uh from this database and to get the list of our articles in wiki data I used a database uh of james here you shared in fiq chair here that is uh extracted using w damper which is a tool that allows the creation of customized damps for wiki data information so I used that I just read in the line I just uh I did the doi and the wiki data items and I am searching for information user requests from the doi or from the api of open citations so thanks to this I have I will get the doi of all references of each publication in wiki data and to uh to reconcile this dois for the references with wiki data items uh actually uh uh found in uh in wiki data I use this tool which is the wiki data hub tool that allows to return the wiki data item of another wiki data item based on its external identifier and hopefully it has uh gson uh database okay so uh I use this one I'm I I got the wiki data items and I have shared I've shared uh somewhere okay so later on if the wiki data item does not exist in wiki data I have to create a new item so hopefully open citations api has the doi uh has a metadata api that can be used to return metadata for uh an an available article so we can get all the data from there and you use them to create new items for in-existing items in wiki data so once all the data is here where we use wiki uh wiki base integrator as a tool to must upload new information to wiki data and wiki base integrator I must recognize that there are many options for uploading the such information to wiki data there is the the media wiki api there is the py wiki bot but I think and I believe that wiki base integrator is the most easy way so all what you have to log in is to create a login instance and then according to the type of the object you have to set the the statements and then create the references like this create the qualifiers and set up all the things and using uh item engine and write you can add directly informations from wiki data so we have done some some sets of uh of uh trials using uh open citations bot and we found that it has been successful indeed to create new items as you see here as well as to create as well to uh to add citation work relations to wiki data so that's all what I have to share I have just to say that this could not be possible without showing it this bot for in the first day of uh of wiki mania apikaton and thanks to interaction we have got plenty of issues that have been solved to get the better output so thank you wiki mania apikaton thank you house and media and this was our last presentation for this session before closing the session I want to thank to all of the contributors who gave their efforts during the hackathon personally I feel full of inspiration after short cases which displays how great projects can be done in really short time so many thanks to participants please do not forget that you can still add your project to showcase page on the wiki even if you don't prefer to present it in live stream and also another kind of reminder shortly we will have a gist meeting for social gathering Q&A discussions and rest of the showcase where you can join without being recorded so I hope seeing you there thanks for listening