Train Visualiser

Station-level rail provision and reliability measures for Great Britain

01Define observations

ORR station calls retain their operator, station and reporting period

02Measure performance

Observed denominators weight cancellation and punctuality rates

03Explore geography

Stations, local authorities, city regions and ITL1 regions link the evidence

04Compare places

Standard measures and exploratory indices support descriptive comparisons

Designed by Nicholas Sweeney

Place profiles

×

Data and methods

Definitions, observed denominators and the limits of the completed analysis.

×
01Define observations

ORR station calls retain their operator, station and reporting period

02Measure performance

Observed denominators weight cancellation and punctuality rates

03Explore geography

Stations, local authorities, city regions and ITL1 regions link the evidence

04Compare places

Standard measures and exploratory indices support descriptive comparisons

01
Scope

The counting unit is a scheduled or recorded station call, scoped to station, operator and reporting period. Operators are summed. Calls are not unique trains. Daily volumes include every calendar day, including weekends.

Stations, LAD24 local authorities, 11 reviewed city regions and 11 ITL1 regions in Great Britain use the same report window. Northern Ireland is outside this rail dataset. City regions use explicit whole-authority memberships; they are not a complete partition of Great Britain.

02
Standard measures

Cancellations: sum of scheduled calls × available operator cancellation rate, divided by scheduled calls with an available rate.

Delays ≥3 min: 100 minus the recorded-call-weighted Time to 3 percentage. It describes calls outside the punctuality threshold; it is not mean lateness in minutes.

Provision: scheduled calls / inclusive calendar days / ONS mid-2024 resident population × 10,000. The population is fixed across the reporting window. A station has no inferred catchment denominator.

Each rate shows its observed denominator and coverage. Estimated disruption counts inherit rounding in the source percentages. Missing data remain unavailable; zero stays zero.

03
Exploratory OCI & ODI

Eligible operator rows require positive scheduled calls, an available cancellation rate and at least 80% recording completeness. Punctuality uses its own available recorded-call denominator.

OCI = eligible cancellation proportion × reference median scheduled-call frequency / eligible scheduled-call frequency.

ODI = eligible delay proportion × reference median recorded-call frequency / eligible recorded-call frequency.

Station pooled frequencies use eligible calendar days. Geographic frequencies use total eligible calls / eligible station-days. Reference medians are calculated separately at station, LAD, city-region and ITL1 levels. Profiles retain the actual medians and eligible sample. Compare indices within a level. Lower scores can reflect higher frequency alongside appreciable disruption.

The atlas uses the report's pooled base indices. It does not substitute operator-network connectivity for direct journey reach. These dimensionless exploratory descriptors are not monetary valuations or causal investment benefits.

04
Geography & provenance

ORR station names are joined exactly to the report outputs. Reviewed TIPLOC mappings and NaPTAN rail references identify station points. Only official station-reference coordinates within the GB coordinate envelope appear on the map; unmatched or invalid stations remain searchable and in tables.

Areas follow the report's LAD24-to-ITL125 lookup and reviewed city-region membership. Display polygons retain the report's 50 m simplification and holes and are transformed from British National Grid easting/northing to WGS84 longitude/latitude.

Boundary provenance remains provisional: the upstream provenance of the supplied LAD24 boundary archive requires confirmation. The atlas is a research preview; polygons should not be used for legal or operational boundary decisions.

Population: ONS mid-2024. NaPTAN reference: 13 October 2026. Timetable weekday/peak calculations use a separate dated window and are not presented as full-window daily frequency here.

Operator source codes [u], [z] and null are retained in station evidence downloads and the operator detail. Missing and non-applicable counts are not converted to zero.