How much a train's promised arrival time changes while riders wait. Smaller revisions are steadier.
FIELD NOTE / DEVELOPMENT SNAPSHOT
Chicago trains,
read between the lines.
A continuously updated reliability observatory that records CTA arrival predictions, measures how those predictions move, and makes service gaps visible.
OBSERVED NETWORK8 / 8 lines
35th-Bronzeville-IITGreen Line01
35th/ArcherOrange Line02
54th/CermakPink Line03
95th/Dan RyanRed Line04
O'HareBlue Line05
KimballBrown Line06
A live dispatch list, not a geographic map. The cohort covers twelve stations and all eight CTA rail lines.
NETWORK PULSE / LATEST STATION SNAPSHOTS
A live reading,
with its limits attached.
Expected wait is the time between CTA's snapshot and predicted arrival. ETA movement compares the same train across repeated snapshots; lower movement means a steadier prediction.
LINEARRIVALSSTATIONSAVG EXPECTED WAITETA MOVEMENT / 14D
Minutes between consecutive predicted trains going the same direction from the same platform.
Predicted gap divided by scheduled headway. A value above 1.0 means the observed gap is larger than planned.
STATION DISPATCH
Where the wait stretched.
01
Green Line35th-Bronzeville-IIT
12.4min
expected wait
expected wait
5 upcoming arrivals · 0% currently marked delayed by CTA.
02
Orange Line35th/Archer
18.9min
expected wait
expected wait
5 upcoming arrivals · 0% currently marked delayed by CTA.
03
Pink Line54th/Cermak
18.4min
expected wait
expected wait
5 upcoming arrivals · 0% currently marked delayed by CTA.
04
Red Line95th/Dan Ryan
12.7min
expected wait
expected wait
5 upcoming arrivals · 0% currently marked delayed by CTA.
05
Blue LineO'Hare
1.7min
expected wait
expected wait
1 upcoming arrivals · 0% currently marked delayed by CTA.
06
Brown LineKimball
0.7min
expected wait
expected wait
1 upcoming arrivals · 0% currently marked delayed by CTA.
“
Reliability is a story over time—not a score from one snapshot.
That sentence is a product rule. Signal Chicago keeps the raw evidence long enough to calculate stability, then preserves compact daily summaries.
THE TWO-MINUTE DATA JOURNEY
01Collect
A protected Supabase function requests CTA predictions for twelve stations.
02Validate
Timestamps, station coverage, flags, and errors are checked before loading.
03Compare
Repeated ETAs become revision and service-gap measurements in PostgreSQL.
04Publish
The public site requests read-only summaries and refreshes once per minute.
FROM THE BUILD LOG
The first cohort balances route coverage, terminals, transfer stations, airport service, and an Illinois Tech connection.
Each successful run writes predictions and an audit record. Partial station failures remain visible instead of silently disappearing.
Fourteen days of raw predictions support stability analysis; daily reliability summaries remain for long-term comparisons.