I've been collecting data from my phone and some different "sensors" (food delivery orders, bank payments, Activity Watch statistics, I'm adding a video camera soon. For phone about every sensor I have in my phone like ambient light, wifi APs, GPS, rotation, steps, when it was last unlocked, what app is in use, what app / artist is playing media etc.) and I'm using ML to predict activities I'm currently doing. I guess the end goal is to automate my diary writing to GPT as a joke..

Started as a small project on wanting to learn C# better and being interested about how much sensors collect information about us and how we could analyze that, but it really got out of hand. I've been working on it for a ~month (with some breaks), sometimes all day, built a lot of tooling (like webassembly & android app to tag current activities for ML training data) and analysis things for it and am nearing a burnout point :P but been fun and I've definitely learned a lot about C# and ML.