Melbourne SCATS Topic Pages
SCATS Data Methodology
This methodology page summarises the processing layer behind the public Melbourne SCATS analysis: cleaned interval observations, deduplicated outputs, monthly chunking, validation checks, chart generation and evidence packaging.
Usable 15-minute observations
37,877,397,311
Cleaned vehicle movements
539,020,710,239
Monthly windows processed
148 / 148
Daily records
4,437
Loaded range
2014-01-01 → 2026-04-07
Known data gap
2018-12
Processing Workflow
Raw SCATS Files
Daily CSV-style source files are ingested into database layers for analysis.
Clean Views
Negative sentinels and invalid values are handled before reporting.
Deduplication
Unified clean views remove duplicate detector-day records where required.
Chunked Processing
Heavy computations are run month-by-month to avoid memory blowouts.
Outputs
CSV, JSON and PNG outputs feed the public HTML report.
Validation and Limitations
- 2018-12 is flagged as a zero-volume data gap, not a real traffic outcome.
- 2026-04 is partial because the archive ends on 2026-04-07.
- A vehicle event is not a unique vehicle; the same vehicle can generate repeated counts over many trips and intersections.
- The page is designed for public intelligence, journalism, planning discussion and OOH screening rather than replacing official statutory modelling.
Methodology Charts
Processing Time by Completed CSV
Shows operational runtime behaviour for completed outputs.
Yearly Daily Percentile Bands
Shows distribution-level changes rather than simple averages.
