Ask your map data a question. Get an answer you can defend.

Data Explorer answers plain-language questions from the map services your organization already runs, and shows the method behind every answer: which dataset, which boundary, which filters.

Open the live demo Why it works this way

Find fatal crashes in 2024 within 1 mile of Fairfax

Which Fairfax did you mean? Fairfax City Fairfax County

Found 62 fatal crashes in 2024 within 1 mile of Fairfax County. Of the 62, 17 involved alcohol and 13 involved an unbelted occupant.

How was this answered?

Location filter: Within 1 mile of Fairfax County (county boundary, buffered)
SQL: (CRASH_YEAR = '2024') AND (CRASH_SEVERITY = 'K')
Source: Crashes, data as of July 20, 2026

A real exchange from the demo. Every answer ends with its record: rerun it next month, or change one part and ask again.

It asks before it assumes

Two places called Fairfax? It offers both and records which one you picked. The area a question covers never changes without telling you.

The method comes with the answer

Dataset, boundary, filters and the time the data was read, in a form a person can check and a machine can rerun. A refusal comes with its reason in the same form.

The AI can't improvise a query

It chooses from operations your administrators approved. Every column and value is checked against your real data before anything runs.

Your experts set the behavior, and can prove it

The person who knows the data decides which dataset answers a question and what each code means, in an admin screen, with no code. Then they ask a real question and watch the change take effect.

Built on what you already run

For state and local agencies, utilities and the firms that work with them: the analyst who answers the same request every week, the records officer who has to stand behind a number, and the IT lead who has to approve it.

Your services: Read-only. Your GIS team keeps publishing as it does today, and the sources stay the system of record.
Your infrastructure: One container, deployed where you approve.
Your choice of model: A commercial service, one in your own cloud, or a private one on your own hardware. Small models are enough.
Your servers' load: Heavy questions run on a local copy that administrators control, so they never compete with your map viewers.
Your people's knowledge: Written down once by the people who hold it, with a name and a date on each decision, and used in every answer after that.

Try it on real data

The demo runs on publicly available map services. Ask it something an analyst would be asked.

Open the live demo [email protected]