Tagged
Everything currently filed under mashup.
2013
I wanted structured data about Coronation Street and Fandom does not have DBpedia. Infoboxes are still in there, so I wrote a Python scraper, then later pointed a Haar cascade at a clip to see if it could tell Emma from Steve.
Soap-opera wikis, infoboxes, a Python crawl, and a face detector pointed at Coronation Street.
2024–ongoing
A web app for ranking work by repeatedly picking which of two things is better. A collaboration between Liverpool’s Centre for Educational Development and Support, the School of Dentistry, and the School of Biosciences. Students use it to peer-assess and actually look at the work; staff use it to mark at scale; it has also been used with movie posters as an icebreaker. Dentistry is using it live.
Keep picking which poster is better, zoom in like a museum, and keep a spreadsheet people actually trust.
2025–ongoing
A local OBS browser overlay where chat plays a ridiculous retro text adventure together: wandering around, voting on directions, fighting bosses, racing, doing pub quizzes, and slowly becoming an adventuring mob with levels and grudges.
Twitch chat driving a text adventure on an OBS overlay, with YouTube chat mixed in.
2021
My PhD. I ran collaborative workshops where people used open-source stats on digital footprints, hoping they’d get a feel for what data-driven organisations actually do with their data. Teaching the methods bored most people. Showing wrestling tag-team graphs to wrestling fans worked better. University of Bolton, 2021.
Open-source stats, digital footprints, and wrestling graphs, trying to make surveillance feel real enough to talk about.
2014
I pulled wrestling stables out of Wikipedia via DBpedia, mapped who had been teammates, and posted the giant graphs on Reddit. Wrestling fans spotted missing Ric Flair, argued about stables versus factions, and the messy Wikipedia data turned out to be the interesting bit.
Wrestling stables, Wikipedia, and a giant network map that Reddit fact-checked.