v1.0.0
First release: a small tool that finds the best time to leave for your commute by measuring it, instead of trusting Google Maps' wide estimate ranges.
Features
- Adaptive polling of the Google Routes API (traffic-aware): every minute during configurable rush windows, every 10 minutes off-peak, optional active window — ~6,600 calls/week for both directions, inside the ~10k/month free tier
- Two-direction tracking in one SQLite database (morning and evening commutes from a single run)
- Weekly report (
report.html): median commute vs. departure time per weekday, plus best-departure tables within each rush window with minutes saved vs. the worst slot - Live watching (
commute watch): regenerates the report plus a day-by-day sub-report mid-run so you can see data before the week is done - Preflight check (
commute check): validates config, API key, database, live API access, and estimates quota usage - Auto-stop timer (
poll -d 7d): ends the run and generates the report automatically - Demo mode to preview the report with a synthetic week of data
- Zero dependencies beyond Python 3.9+; works on Windows, macOS, and Linux
Known limitations
- Polling pauses if the machine sleeps or reboots (no auto-restart; disable sleep for the run)
- Multi-week data blends by weekday — use one week per database
- API errors are logged but not retried with backoff
- No packaging/tests; run from the repo with
python -m commute
🤖 Generated with Claude Code