On the last day I checked and yes, I was the southern-most Wikimanian at Poland as well as in Singapore!
Wikimania 2024 ran from 7–10 August in Katowice, Poland. I was lucky enough to have been put on the reserve list for a scholarship, and was offered a scholarship a couple of weeks beforehand. It was a bit of a scramble to arrange, but I made it! These are my notes about what I did and who I met, and thoughts to follow up on. I did propose a talk, but it wasn't accepted.
The day before the conference started, there was a full day GLAM workshop at the Silesian Museum. While it was aimed primarily at GLAM professionals, I work with library metadata for the NZThesis Project, so was keen to talk to other library people. We talked a lot about the tools we use to do our work, and at the end had an 'auction' of projects people would like to collaborate on. I volunteered to help out on Lucy Moore's smell project and Martin Poulter's improve-a-neglected-Wikiproject work. I was pleased to meet Martin, and talk to him about thesis work, and had some valuable feedback from Annika about my Mix'n'match video (I need to redo it with better sound!). I also chatted with Kasia, Peter who is interested in uploading theses from his library in Australia, and Tove who is also thinking about extracting theses from DSpace, and Stuart, who uses some of Martin's queries to show how things are connected on Wikidata. One very useful thing the organisers of the meet-up did was create a slide deck for participants to upload their image and a short description of themselves before the workshop, so we didn't have to spend a lot of time introducing ourselves.
Opening ceremony - unfortunately much too long. I enjoyed the cooler temperature of the room and really appreciated the excellent air quality as shown by my Aranet readings. Wikimedian of the Year awards are a highlight of this session.
10.35 Media education, libraries and schools - Czech experience. A session about how the presenters had been teaching Wikipedia editing to highschoolers as a basis for media literacy. Included an exercise on recognising text sources. They've written a handbook for teachers with five practical activities to do with students (handy to know what they try to cover in a 1 hour or 45 minute lesson). Also included discussion of Czech upload of library catalogue to Wikidata. Afterwards mentioned to Lucie that in the thesis project we have more Czech library identifiers linked than NZ ones, she was interested to get a quote about that, need to follow up. Session slides.
1.30 Museums, Media, Data & Gender: How to Make Images of Women More Visible on Wikimedia? I think jetlag and bad ventilation took a toll and I don't recall a lot of this session, although Lucy Moore's work in Leeds was memorable. Probably need to go back and view the recording! eventyay link I did note that the way images spread through Wikis quite easily and fast means they're quite an 'economical' thing for Wikimedians to work on, with potentially quite a lot of bang for buck.
4pm Open as a prerequisite for solving the climate crisis. Useful as it got me thinking about how to use organization's own climate goals as leverage for getting them to open up publications. Convinced me even more that we need to try to get our entire repository uploaded to Wikidata. eventyay
Keynote "Opening the academia" Missable. And I was annoyed that one panellist was saying journals should collect editor names for authors. It seems obvious to me that collecting Wikidata QIDs (that link to editor names, Orcids, Google, etc etc) is a much more obvious thing to do and applicable to many many more authors.
Wikidata Query Service - the way forward for getting the most out of Wikimedia's knowledge graph. The Graph Split. Who knows when, but at least we have been told yet again why. I am not convinced that it will take us two years to fill Wikidata up again, but let's accept the challenge. I at least feel that we can take the brakes off scholarly publication imports and citations now, right?
Smell on Wikidata - Lucy Moore and Glory did an excellent session on smell and how we model it, which is really interesting but also has health implications. Tech problems meant we didn't get the demonstration from Glory about how she models smells in relation to health, but I hope she can record something to be shown another time as it seemed really important. Follow-up: created the Wikidata Wikiproject Smell.
Higher education meetup. This wasn't a meetup for people in or interested in, higher ed, but people teaching Wiki in higher ed, ie not much relevance to people interested in academic libraries, university collections etc. (Me: Why are the library, higher ed, open access interest groups so separate?)
Demolishing the ivory towers: bringing universities and open knowledge communities together. In which I declared my objective of getting the world's theses into Wikidata. Interested in creating a funding proposal for a network grant. I was the room manager for this, and the three following Spanish talks.
Alumnos de Secundaria, autores de 1.000 artículos en Txikipedia Saw this on in English in Singapore, I think, or online. Hilariously I discussed with the notetaker Clover beforehand and found that neither of us spoke Spanish, so she asked to switch sessions - and was replaced by a notetaker who didn't speak Spanish! Thank goodness for excellent live translators, and someone in the room who translated an eventyay question for me. Also one of the first speakers went off with the slide clicker in his pocket and had to be tracked down, that was funny.
Métricas sobre el bienestar de nuestro movimiento. Ampliando las colaboraciones con otras comunidades. Metrics? I tried listening to the English translation but had to stop as they wanted the microphone running the room during the session. There was no way for me to communicate with the translator about the speed, I suspect it was too fast.
How did the Spanish LGBT Wikiproject go from 5 to 750 new articles per year? (This was originally going to be in Spanish but was switched to English). Inspiring work. Not normally a fan of using shame as a motivator but the icon of Barbot who "will judge you if you don’t help close the LGBT content gap" was hilarious and very cute, and clearly had the desired effect.
Introducing the Citation Watchlist. Useful idea not sure if I'll use it myself though. eventyay
Universities and Wikimedia: Increasing Participation through Student-Run Efforts. A student editor talking about why the university-run editathons at his uni library were not successful and how they could have been. (Upshot: get a student club going). Eventyay
Student engagement with openness. Excellent overview of years of initiatives at Edinburgh. Now we just have to get their theses in! Eventyay
Had an excellent corridor conversation with Lydia Pintscher about Wikidata and the graph split. I showed her the Scholia button in use on enWp, which she was not aware of, and Wikidata calls in infoboxes. Discussed the potential for a Wikibase for publications, the graph split, the need for a list of tools which may break when the split happens, and possible support for tool developers to rewrite queries. Lydia later reported that a page had been made for this (I just have to find it and see if it has been publicised anywhere!) Note: there is a way somewhere to check how many Wp pages a Scholia template is used on.
Wikimedia and libraries meetup. We spent most of the time introducing ourselves, which rather limited opportunities to get beyond that. We did hear that the ExLibris Primo system now pulls from #Wikidata if the item uses the Wikidata property "archives at", so should follow up to see how widely we are using that property in NZ, and whether we can get all libraries to submit who they hold the archives for. Also as a non-librarian I don't know who is using the Primo system, but it would nice to see it pulling the data.
WikiPortraits. Getting celebrity photos at big film festivals but also potentially with useful workflows for other events, e.g. we should be doing a lot of academic conferences in similar ways. Would love to see this project produce documentation on how to (so I don't have to reinvent the wheel when I volunteer to collect the photo metadata). Also Andrew Lih explained his Google Sheet that uses the Wikipedia API to fill in a heap of information. Andrew told me that because I told him last year in Singapore that I had used this sheet, he took the time to include it and explain it today. That was lovely to hear!
OpenRefine meetup. Being a bit of a newbie to meetups at Wikimania, I didn't realise that in requesting to have the meetup, I should perhaps then have organised an agenda and be responsible for what happened! Guided by Antonin, we raised the topics we were interested in and split into discussion groups around education on OpenRefine, and qualifiers (what asking for qualifiers would look like, how to arrange information when it's extended) and ranks. Showed someone how to use 'fill down'. Next time it would be great to find topics of interest to the group beforehand, and those things Antonin would like to hear from the community about, so we can get the most out of our in-person time.
My feet got quite sore, I took off my shoes and walked around in my Wikimania socks for one of the days.Equitable approaches to hybrid events. Raised the issue of also surveying the people that didn't attend your event, or you won't find out why they couldn't. Time zone poverty etc.
Tool sustainability meetup. Lots of technical discussion, Mike and I trying to say that Toolhub/Toolforge/Phabricator etc are not places tool USERS hang out and as users and not developers we need better ways of knowing what tools are supported, where to report errors and so on. It is very frustrating to learn about a new tool at the end of its development and then discover there is no support for the users (or developer) when it goes wrong/needs adjusting to fit a browser update etc, a few months later. Heard that Wiki Brazil are developing an alternative to QuickStatements which is interesting.
Open Space/Future of Wikipedian in Residence position. A lot of discussion around all the work that needs to happen before a position gets established, and also the despair that follows when a WiR leaves and the work all stops, how to create a lasting effect? Also shared concerns around conflicts of interest etc.
Closing/awards. Is it just me or is it odd that a country can send more than a hundred Wikimedians and there is noone to stand up for that country at the closing ceremony? Maybe they all went on a tour together. A lot of information from the organising committee seemed repeated from the opening, which was unnecessary and slowed the pace. The Cool Tool Awards were a highlight of course, although I was relieved not to have be on stage when my video played! Having spent some time running around trying to get my photo taken for being on the Cool Tool Academy, I was amused that none of the photos seemed to be used. Link to my video about Web2Cit.
Wikimania Party. There was meant to be dinner, but the food served had two tiny vegetarian options and I was left starving, which was a shame after what was otherwise a well-catered conference. The fashion show was fun, as was chatting with Annie Rauwerda and Adam. I left earlyish as it wasn't a very Covid-safe environment, and then spent another hour chatting outside about authority control with Epidosis, including tricks for getting conflated VIAFs resolved.
Wikimania post-conference Sun 11th August, other conversations
German theses conversation with Sabine, pointed to Heidi's list of German botanical theses
On a number of different days: Long conversations with Liam Wyatt re WikiCite, and Alessandro and Camillo re authority control and WikiCite. Daniel Mietchen re Scholia (learned a trick to find which Orcid profile a pub is cited in), pubs in general.
Discussed Web2Cit worklist with Mike, needs to be reviewed, updated, a plan for Stuff/PapersPast figured out (could a Zotero template fix either of these issues?)
10 Research Findings and How You Can Use Them in Your Work (eventyay has links and slides).
Biases in makeup of editors persist. We are disproportionately young, BUT 93% of admins are male! Sigh.
Survey of readability says 58% of enWiki articles are set at a higher reading level than that of the average American. (Readability of an article is not something our quality tools currently assess, but the readability tool on toolforge is available for testing. I can't get the button to do anything though!)
15% of articles are orphans, and women's biographies are overrepresented in orphans. De-orphaning leads to +10% persistent readership of an article. But do-orphaning is not often included in training for new editors. Having a de-orphaning event/campaign could be useful.
AI. An mBERT model developed for writing for writing article descriptions. Now live on Android app. Multilingual Revert Risk model developed, available, for Wikipedia in 47 languages; developed by not yet available for Wikidata. Link recommendation model improved.
Models generating text produce many wrong answers as well as some right ones, so output must be checked by humans.
Gathering the data is difficult - labelling of data (e.g. edits that got reverted) is more difficult and still depends on volunteer effort. No system to bring labelled edits back to research team.
There is an infrastructure cap, using link recommendation on 300 different Wp used up the entire capacity.
Wikidata and authority control for researchers: the cases of Switzerland and Italy (eventyay for links and slides)
Approaches the contribution of researcher items to Wikidata from the perspective of having authorities for researchers e.g. government database. Achieving similar things to the thesis project from a completely different angle.
.png)
