v4

latestOpenAPI 3.1.0Proprietary2026-08-034715226.2 KB
analytics

Incident-probability forecast

Empirical incident probability and expected delay for a road segment by day-of-week and hour, derived from historical aggregates. Omit dow/hour to get the full 7x24 grid. Thin segments degrade to corridor then jurisdiction basis; sample_size, confidence and basis report how much data backs each bucket. Requires Pro+ plan. expected_delay_min and p95_delay_min count only unplanned event types (incident, closure, weather) — a three-week work zone is reported by probability, not as a three-week delay — and each contributing event is capped at 4320 minutes (72 hours). A value at 4320 therefore means "at least 72 hours", not exactly 72. Both fields are omitted when no duration could be established; the event still counts toward probability and expected_count.

get/analytics/forecast

Query parameters

jurisdictionstring required

Jurisdiction code (e.g. GA).

roadstring required

Road name (e.g. I-85).

road_classstring

Optional road class filter.

dowinteger

Day of week (0=Sunday..6=Saturday). Omit for the full grid.

hourinteger

Hour of day (0-23). Omit for the full grid.

Response

Segment forecast — a single bucket when dow+hour are pinned, otherwise the full grid in buckets[]