Urban Road User Counts from AI Sensors in 5-Minute Intervals

Hourly-updated traffic activity data from AI-based AIRS sensors, detecting 14 road user types including pedestrian, cyclist, e-scooter, and multiple vehicle classes. Historical data is available from 2023 onwards, packaged in quarterly ZIP files. Readings are provided in 5-minute intervals, though not every road user type is reported for every interval.

Use Cases

• Model hourly and daily transport activity patterns for road user types like Car, Bus, and Cyclist using 5-minute interval data.
• Analyze modal share shifts by comparing time-series counts of Pedestrian, E-scooter, and Motorbike across sensor locations.
• Detect anomalous traffic events by monitoring spikes or drops in counts for specific classes such as Emergency Car or Truck.
• Forecast short-term demand for infrastructure by training on sequences of Van, Taxi, and Rigid vehicle counts.

Strengths

• Data covers 14 distinct road user types, enabling detailed modal analysis.
• Updated hourly with historical records available from 2023, providing recent temporal coverage.
• Collected via AI-based AIRS sensors, offering autonomous detection and classification.

Limitations

• Data may contain gaps, as noted, with no 5-minute reading guaranteed for every road user type at every interval.
• Unknown total row count and dataset size limits pre-assessment of statistical power for some analyses.
• Spatial coverage is limited to sensor locations across a single city, not representative of broader regions.

Provenance

Source

City of Melbourne Open Data

Collection Method

Collected by Artificial Intelligence Road Survey (AIRS) sensor devices that autonomously detect and classify road users.

Time Range

From 2023 onwards

Freshness

Updated hourly; last updated 2026-03-20.

Geography

Sensor locations across the city (likely Melbourne, Australia)