Picking a first project when you have no ideas

Can I finish a first version of this in two weekends? Does it involve something I can actually measure, build, write, or show? Would I be a little proud to send it to someone I respect?

If a project idea passes all three, it's good enough. Ideas that fail 'two weekends' almost always fail entirely — you'll lose momentum before you finish.

Pick a public dataset (Kaggle, data.gov, a city open-data portal) and ask one specific question about it. 'Which subway stations get most crowded on Saturdays' is a whole project.

Replicate a small figure from a paper you liked. Not the whole paper — one figure. You'll learn more than you'd expect.

Solve a real annoyance in your own life with a small tool, script, essay, or study. The specificity is the whole point.

Interview five people about a problem you find interesting. Write up what you heard honestly. This is a real project, and it's undervalued.

If a project feels exciting but overwhelming, ask 'what's the smallest version of this that would still be interesting?' Cut features. Cut scope. Cut ambitions. Ship the tiny version. Then extend.

A finished small thing is worth ten unfinished ambitious things. Every experienced person knows this. Very few students believe it in time.

Give a new project two full weekends. If after that you still can't articulate what you're doing and why it's a little interesting, let it go without guilt.

Abandoning a bad project early is a skill, not a failure. Sunk-cost thinking is what turns two lost weekends into a lost semester.