I have a bit of a obsession with the Lisbon Metro, even though I don't use it very often. I started by building this:
https://github.com/gbl08ma/disturbancesmlx
This started as more of a statistics page for the service interruptions published by the Metro on their website, which I scrape. The slight tongue-in-cheekiness of my website, which opens up with a large text saying something like "XX days since the last disturbance", where XX is usually a single-digit number, made it become mildly popular (at least in terms of what I'm used to).
This particular subway system doesn't operate on a fixed schedule and doesn't show the ETA for the next train outside of the platforms nor on any app or website. (Google thinks there's a schedule, but they've been fooled.) They also don't publish usage stats for each station, which would be of great interest to everyone who likes daydreaming of expansions, network reorganization and the like. Furthermore, I read and heard multiple reports of delays and interruptions that never made their way to the website. So I decided to build a Android app to unobtrusively crowd-source data and communicate the service status back to the users...
https://github.com/gbl08ma/underlx
...and the very ambitious goal is to, one day, be able to calculate train positions and ETAs based on real-time data reported by the smartphones of people riding the subway. Pretty much "Waze for the Lisbon Metro".
Yeah, I've put months of work into this and there's absolutely no business plan; it perpetually feels 5% complete. But it's been fun putting together my second Android app, playing around with Postgres (after many years using MySQL), designing the REST API and writing the server in Go. I plan to use this big project as my sandbox for experimenting with machine learning and other AI techniques, as well as data analytics and visualization. There's already a small but extremely interested group of users, which really motivates me to keep working on this.