My (mostly) finished projects
This is a list of projects that I’ve published, either because I consider them good enough or because I lost interest in them but was happy with the result, even if partial.
Kindertip
It’s an alternative frontend for the app (Kindertap) used by the kindergarten where my daughter used to go, and where my son will go soon. The app is used to give parents information on the food eaten by the children, and share pictures.
It uses the same endpoints as the app, but it focuses only on the food, showing historical data and highlighting favourite meals.
I found very useful being able to understand with numbers which meals I could replicate at home.
Magic Top Hat
This is a small python project: it’s a library that uses an llm to generate a function at runtime based on the context in which the function was called.
You can import any kind of function from this library, with any name, and once you call the function the llm will be asked to generate some working code that solves what the function seems to need to do.
It’s a joke, it’s of course useless if not dangerous, but it was really funny to write this. Also, thanks to python magic, when you try to introspect the library and check what’s inside, it’s empty.
Olympic medals weights
This was another short project: during the Paris Olympic games of 2024 I wanted to check how much some medals were correlated with each other, giving a bigger weight to medals that seldom, or never, had winners that also won any other medal.
Getting the data and the results was quite simple, then I wanted to create a simple viz that resembled medals: that took more time than the rest of the process.
Also, I found some interesting Trivia.
Manifestival
This is a project that I did not finish in time, and will likely finish at a better time. In occasion of the 2024 european elections, I used two agents built using an LLM to impersonate the manifest of a party each. They were then asked a question (such as how to fix the economy) and they had to discuss.
My goal was to then create a frontend where users could choose two parties, like in a fighting videogame, and they would play the prerecorded discussion.
I never got to do the frontend, since I got the idea a couple of days before the election, but I’ll likely make something similar for the next italian ones.
Weighted Voronoi
The goal is to generate a Voronoi where the size of cells is predetermined. I also wanted it to be a dynamic representation.
I first used different minimization algorithms to find solutions, but then added some noise so that the movement seemed more organic.
I think it looks nice, but not enough to put real data to show.