This week's dataset was on every country's scores and medals in the International Math Olympiad. I wanted to visualize both overall medal count and how rankings have fluctuated over the years. The small multiples are actually bump charts! Perhaps a little overwhelming, but I enjoyed using the {patchwork} package.

After a long time away from Tidytuesday, I made a fairly straightforward graphic showing how the 'prominence' of characters changes as the three plays go from Act I to Act V. Represented with a bump chart, the characters are ranked by how much dialogue they have in that particular act and scene.

I mapped addresses of U.S polling places and plotted their locations (probably a population map lol). Geocoding was an interesting challenge; I found out that the Census website provides a nice generous_API. Processed in R but mapped in QGIS (whoops)._A pretty simple visualization for this week, but getting geocoding to work was fun. It took around 4 hours to run through the full dataset in batches of 9999 addresses each.\n

My friend and I are hyper-competitive with our NYT word games, this year we want to see who is *actually* better. Using this week's #tidytuesday prompt to create a Shiny dashboard that tracks our daily plays, wins and win margins for the rest of the year. The game is on

It's tidytuesday week 48 and we looked at Doctor Who. I made a little nerd postcard from Gallifrey. I think I got the shape pretty close? eh?? Added the unstylized versions at the bottom. Except the title graphic, everything done in R.

For #tidytuesday week 47, we looked at @RLadiesGlobal chapters, and I probably overdid things. Tbh, I just wanted to use the OpenAI embeddings API to cluster similar chapters but one thing led to another and well yeah, I made a dashboard. You can search by topic or locations. All the processing for the data was done in R and the website is (obviously) built with Sveltekit. I used the OpenAI APIs to extract keywords for tagging and to 'recommend' chapters based on the kind of events they host. It cost me ~$0.90.

For #Tidytuesday week 46, I used weighted log odds to see what kind of products are potentially more 'popular' during Diwali season. Made in r + Figma

For tidytuesday week 45, I accepted that the US election system is beyond my understanding lol. Made a couple of things, but settled on trying to recreate an old @FiveThirtyEight analysis and graphic.

I analyzed repetition of lyrics in Taylor Swift songs. If you take away similar phrases and words from the 1989 album, how many songs can be reduced by at least 60%? All of them! The repetitions make them catchy as heck.