Turning a crash coordinate into an official location — automatically
A plain-language look at what this tool does, why it matters, and how well it's been shown to work.
The problem
When a crash is reported, the responding officer's GPS coordinate isn't the final answer. That coordinate has to be converted into a standardized location reference (which highway and mile marker; or for local roads, which street and how far from the nearest intersection or landmark) before the record is actually usable for analysis, safety planning, or public reporting.
Today, that conversion is often slow to come back and light on detail. It typically returns just the location itself — not the broader set of roadway data (speed limit, traffic volume, functional classification, and more) that real, data-driven safety analysis depends on.
What this does
Provide a crash coordinate and a case number, and the service returns the same kind of official reference-post answer used today — in about a second. Crucially, it pairs that answer with a full roadway dataset that goes well beyond just the position itself.
How well it works
Measured by comparing this tool's output against thousands of the state's own already-finalized 2023 crash records — not a guess, a direct comparison to known-correct answers.
On local and city streets — a much harder problem nationally, since a statewide local-road reference system exists but isn't publicly accessible — following official MMUCC guidelines, the tool correctly identifies the situation (intersection vs. mid-block, on/under a bridge, etc.) roughly 4 times out of 5 in the state's largest cities, and reliably produces a usable location everywhere else, even where the public map data isn't a perfect word-for-word match to the exact recorded phrasing.
Honest about the gap
Rural county roads are the hardest case — some are recorded internally using a grid/mile-number system (e.g. "Road 60"), while public map data uses a completely different street name for the same physical road (e.g. "South 60th Road"). No amount of matching logic can fix two different naming systems describing the same physical road; that gap can only be closed with access to the state's own internal road-naming records. That access — not available publicly — is the single thing standing between "very good" and accuracy expected to approach what's already shown on state highways, applied everywhere in the state. (That expectation is untested against real local-road data today — a claim to prove, not one we're making yet.)
What this could mean day to day
- Faster turnaround — instant answers for the vast majority of crashes.
- Consistency — the same rules applied the same way every time. Every source and assumption is logged programmatically, removing reliance on tribal knowledge.
- Richer output — a full roadway dataset alongside every location answer, not just a reference post: all 18 of the federally-required MIRE elements (23 CFR § 924.17), plus 2 more beyond that minimum.
- A foundation to build on — once proven out, the same automated approach extends cleanly to bulk reprocessing, historical data, or other state location-resolution needs.