I've been thinking recently about crowdsourcing our own dataset of airplane data. When you get on a plane, tap a button that says 'we're delayed 5 mins'. When you land, tap a button that says you've landed. In fact, you probably don't even need users to tap those buttons. The better way would be to use the accelerometer and barometer to automatically detect takeoff and landing.

Airplane data sources are ALWAYS wrong in my experience. I have been on a flight that was 2 hours late but all the flight data sources/apps would say 'On time'. Bullshit. And very often there are delays that the flight data does not account for, but are real delays. User's phones could provide a much more real-time view on where airplanes actually are than any existing flight data source. I hate sitting on a plane and knowing that I am more up-to-date on how late we are than the official data source is. That's super annoying and it seems like there's an 'easy' fix: just build a new kind of crowdsourced, real-time flight data service.

I've seen enough cases of people creating crappy free "replacements" for premium apps or services, and people flocking to them, that I don't know if you'd actually need to fully replicate the data sources. Just find something "good enough" and people will come. Or pay for a little while, planning to "figure out monetization later" and then fail in six months or a year.

That just means that the free alternatives won't stick around, but you could still get a steady stream of people who see your prices, assume they can undercut you and still make a profit and take away a few months of revenue each before they run out of money and give up.

They're welcome to try, but you're correct.