Headway Reality Check

Schedule vs. Reality

CTA schedules 97–100% of Frequent Network headways at 10 minutes or less. This page tracks how often buses actually arrive within that promise, based on real-time vehicle positions collected every 5 minutes. Collection ended March 30, 2026.

Phase 1 Phase 2 Phase 3 Phase 4

How We Measured

We poll CTA's Bus Tracker API for real-time vehicle positions on all 20 Frequent Network routes. Stop arrivals are detected using a crossing+interpolation algorithm: when a vehicle's distance along route (pdist) crosses a reference point (selected at the midpoint of each route), the exact arrival time is linearly interpolated between the two bounding observations. Headways are computed as the time between consecutive arrivals at the same reference point traveling in the same direction.

Scheduled headways

Scheduled headways come from the CTA's GTFS static feed, which provides planned stop times for each trip. We computed the time between consecutive scheduled arrivals at the same reference stop and filtered to the Frequent Network service window.

Service window

The Frequent Network promises 10-minute headways during specific hours:

Weekdays: 6:00 AM – 9:00 PM
Weekends: 9:00 AM – 9:00 PM

Both scheduled and observed headways are filtered to these windows for comparison.

Data collection

Vehicle positions were collected every 5 minutes via a Cloudflare Worker and stored in Cloudflare D1 from February 16 to March 30, 2026.

Limitations

Frozen dataset. Data was collected every 5 minutes during service hours via a Cloudflare Worker between Feb 16 and Mar 30, 2026. Earlier results may be based on limited samples. At minimum, two weeks of continuous data across weekdays and weekends are needed for robust conclusions.

Reference stop selection. We use a single reference stop per route (the midpoint). Headway performance may vary at different points along a route — buses can bunch or spread out as they travel.

Arrival detection. The crossing+interpolation algorithm detects arrivals when a vehicle's distance along route crosses a reference point between consecutive 5-minute observations. A 30-minute minimum gap prevents false duplicates from GPS jitter, but may occasionally merge closely-spaced legitimate arrivals.