The Hitchiker's guide to Open Source

"Solve your own problems" is the oldest advice in open source. But most people think that their problems aren't big enough, complex enough, or interesting enough to build something around. That instinct is probably wrong.

I recently moved to Delhi, and have been trying to attend more events in my quest to become more social. The city has tens of events happening almost every day, and while that is brilliant, it is overwhelming and almost impossible to keep track of everything. Event details and registration links are spread across websites, event aggregators, whatsapp groups, Instagram pages, mailing lists and whatnot!

dilli.today (Currently WIP) is an open-source project that curates events happening in Delhi. It fetches events from multiple sources, cleans them up, then curates them further by tagging them nicely, and makes all event data available as calendars you can subscribe to. It is inspired by and a (vibe-coded) fork of the blr.today project by Nemo. You can find the thesis, roadmap and history of the project at blr.today/about.

dilli.today is a fork and retains the original licenses and notices, and any modifications in this fork continue under the same copyleft terms.

The stack is split into three parts -

ingest - ingestion pipeline that scrapes and normalizes event data

dataset - the dilli.today dataset updated daily

website - A jekyll based website that shows curated events.

To the best of my ability (as a non-techie), I'll explain the technical architecture of the project, the "art" of writing scrapers, deployment setup etc. I'll also explain how exactly I went about "forking" the project and customising the codebase to my usecase (and city) - to the point where I am now able ot build features of my own that I can hopefully contribute upstream.

We'll talk about why "small problems" are underrated as a starting point, and how hitchhiking on existing projects can be a superpower.