This summer, Fatboy Slim released a new single, Role Model. In the music video, you can see many of his role models appear. I’ve been a fan of his music for a long time, especially since he started releasing weekly episodes of Everybody Loves a Mixtape during the lockdowns in 2020.
After listening to this song a lot over the summer, I decided to research who all the people featured in the video are. During that process I created a data model on Wikidata about the music video. In theory, AI assistants like Alexa or Google Voice could now list all the celebrities featured, as Wikidata, like Wikipedia, is free for everyone to use.1
Thanks to this data, you can now do some really fun, nerdy things, like create a chart of the birth years of the people appearing in the video:
It’s not surprising that most of Fatboy Slim’s (born in 1963) role models come from the baby boomer and Generation X generations, but almost every decade of the 20th century is represented. The only missing decade is the 1980s, when I was born. You can also create a map of where these people were born:
The oldest person is Salvador Dalí, and the youngest is Rita Ora. She is the first woman to appear in the video and was unfamiliar to me. Fortunately, with Google Lens, you can quickly match names to faces. Ora is also the one born the furthest east, in Kosovo. The only person not born in Europe or the United States is Bob Marley, Jamaica’s most famous icon. William Shatner was born in Canada, but is featured in his Captain Kirk costume. Star Trek is one of the few domains of nerdom I’ve never gotten into, but even I know Kirk is going to be born in Iowa.
To split the video into different scenes, I used PySceneDetect, software you can control via the command-line interface. Without this, I would likely have missed some people who appear only briefly in the fastest montage of the video.
I didn’t recognize Spike Jonze myself, and Google Lens thought I wanted to know more about men in black suits. For him and three others I couldn’t identify, I sought help from the Wikidata Telegram Group, which is the place to turn to when you’re stuck with Wikidata.
I enjoy the song even more now that I can name everyone. Soon, I plan to watch the experimental 1967 film Herostratus. This is the source of the footage featuring Helen Mirren, one of only 17% women on the list.
With ChatGPT, even non-experts can quickly retrieve more data from Wikidata by simply describing what they want to know in natural language. The generated SPARQL query (pronounced ‘sparkle’) can then be copied into the Wikidata Query Service. The item number for the Role Model music video is Q130298085, where the “depicts” (P180) value is defined. For example, the following prompt retrieves an overview of the people in the list with their photos:
Give me a SPARQL query for an overview of photos of those depicted (P180) in item Q130298085, using #defaultView:ImageGrid, one photo per person
I went though a few more detailed promts to get to get exactly what I wanted, and eventually got it to generate a query that gives the following result:
As an autistic person, I love diving deeply into topics and discovering new things. For me, this goes beyond personal interest; by adding the data I collect and process to Wikimedia projects, others can benefit from it as well.
Footnotes
- Simonite, Tom Inside the Alexa-Friendly World of Wikidata Wired Magazine (19 February 2019)
↩︎
