CTA Bus Tracker, Train Tracker & the Developer Feeds
The CTA started tracking buses on one route, #20 Madison, on 5 August 2006, covered the fleet by May 2009 and added Train Tracker in January 2011. Riders now meet the predictions largely inside other people's apps: the agency publishes Bus Tracker, Train Tracker, alert and GTFS-Realtime feeds, each behind a free registered key. The licence is not an open one — revocable, limited to uses that assist transit riders, and barring anyone from selling the data apart from their app. The feed's sharpest use was against the CTA itself: volunteers scraped it from May 2022 to count buses that were scheduled and never ran. Since May 2025 the agency publishes its cancelled runs, so apps can strike them out.
A peer-reviewed study of route-level ridership across the staged 2006-2009 rollout found Bus Tracker did raise CTA bus ridership, though 'the average increase is modest' (Tang & Thakuriah, Transportation Research Part C, 2012). The CTA reported running 98.8% of scheduled buses and 88% of scheduled trains in June 2025.
🎓 Lesson
Opening the feed let riders audit the service with it: the same data that promised a bus was used to prove the bus never came, and the agency ended up publishing its own cancellations. The licence, however, is a bespoke revocable grant, not an open one.
Evidence🔬 Peer-reviewed
Scale🏙️ City-wide
Sources
↗ CTA Developer Center — open Chicago transit data (operator)Lists the Bus Tracker API, Train Tracker API, Customer Alerts API, GTFS schedule data and the GTFS-Realtime feeds, each requiring a free registered developer key. Direct fetch from this environment returned 403; contents confirmed through the indexed CTA sub-pages, incl. the Train Tracker key application form. The Train Tracker developer guide (PDF, not renderable here) is the origin of the circulating default limit of 100,000 transactions per day — deliberately not asserted in the entry.
↗ CTA Developer License Agreement and Terms of Use (operator)The licence, verified verbatim from a mirrored copy at openmobilitydata-data.s3-us-west-1.amazonaws.com/public/feeds/chicago-transit-authority/165/20150206/original/developers_license_agreement.htm: a 'limited, non-exclusive, non-assignable, non-transferable and revocable license'; data 'may not be used, in whole or in part, for any other purpose' than assisting riders or promoting public transportation; 'You will not sell, auction or barter any CTA Data separate from your application.' Attribution is explicitly optional. Data is supplied 'AS IS' with no representation as to accuracy or completeness.
↗ Tang & Thakuriah — Ridership effects of a real-time bus information system: a case study in the City of Chicago (peer-reviewed)Transportation Research Part C 22 (2012), 146-161. Uses the staged 2006-2009 rollout as a natural experiment on route-level monthly weekday ridership, January 2002 to December 2010. Abstract read at worldtransitresearch.info/research/4150/: 'the provision of Bus Tracker service does increase CTA bus ridership, although the average increase is modest.' The per-route trip figures circulating in secondary reports were not confirmable against the paper and are not used.
↗ Chi Hack Night — Ghost Buses projectCivic-tech group formed spring 2022, project page November 2022. Scrapes the CTA Bus Tracker getvehicles feed every five minutes and compares it with the CTA's own GTFS schedule; collection began 20 May 2022. The clearest case in the atlas of an agency feed used to audit the agency.
↗ Mansueto Institute for Urban Innovation (University of Chicago) & Chicago Tribune — CTA Stop WatchIndependent analysis of over 100 million real-time bus locations from the Ghost Buses archive, June 2022 to July 2024: 'The CTA initially decreased scheduled service in early 2022 through 2023 to match the slower real-time performance and later increased it in early 2024.' Metrics include excess wait time and actual versus scheduled buses by community area.
↗ Grid Chicago — The state of transit trackers in Chicago (2011)Chronology: pilot 5 August 2006 on route 20/Madison, thirteen routes by March 2008, last routes May 2009, Train Tracker 8 January 2011. Also records the sequence behind the API — a Chicago developer reverse-engineered the Bus Tracker feeds before the CTA formally released one. No launch date for the official API is given here or in any source found, so none is claimed.
↗ US DOT ITS Knowledge Resources — CTA bus arrival prediction accuracySummarising the FTA-sponsored 'Traveler Information Systems and Wayfinding Technologies in Transit Systems' (2 May 2011): 'As a result, the Chicago Transit Authority (CTA) has a 92% accuracy rate for bus arrival prediction.' Self-reported by the CTA and relayed by a federal state-of-practice review, not an independent audit — recorded here rather than in the impact field. No independent evaluation of CTA prediction accuracy was found.
1001 Smart Cities (2026). “CTA Bus Tracker, Train Tracker & the Developer Feeds” — Chicago, United States. The Smart City Atlas. https://1001smartcities.org/projects/chicago-cta-tracker/ (last verified 2026-07-26). Data: CC BY 4.0.
The underlying data is free to reuse with attribution — see the open data page.
