About Who makes it, where the answers come from, and how to correct one.

districtry answers one question: which civic districts contain the point you picked, and who represents you there. It is free, needs no account, and is built and maintained by one person.

It began as a Chicago-only explorer and now answers in 6 places. Every boundary comes from the government that draws it and every name from the body that elects or appoints the person, each cited on the card that shows it.

Where it answers

Each place has its own sources page listing every layer, what it answers, and the publisher its boundary comes from.

What it publishes

• 188 per-county pages naming 2,915 county officeholders, one page per county board.
• 1,765 more across the 27 rosters tabled on its question pages, each one dated with the day it describes.
• 128 scheduled jobs that re-read a published source on their own timetable. A job that finds a different officeholder opens a pull request rather than committing, so nothing about a person ships without someone looking at it first.
• 145 recorded data gaps — places where a publisher withholds something, written down in the app rather than papered over.

How it decides what to say

An officeholder is never guessed. Where no verifiable roster exists the card links the official body instead of inventing a name, and says so. Where two sources disagree, the disagreement is recorded rather than resolved by preference.

A boundary comes from the authority that draws it — a county GIS, a state redistricting filing, the Census Bureau, a clerk's certified election returns. Where a boundary had to be derived, the derivation and its checks are written down in the script that did it, and that script is public.

Absence is published. A county whose board this project cannot name has a gap record saying what was tried and what answered, and that record is in the app's own data-gaps panel — not only in a document nobody reads.

Corrections

If a name, a boundary or a contact detail is wrong, say so and it gets fixed. Every correction gets a reply. The fastest route is corrections@overberg.co; an issue on GitHub works too, and the whole history of every file is there to read.

For machines

A short guide for a client reading the site rather than crawling it is at /llms.txt. robots.txt disallows no crawler by name and carries Content-Signal: ai-input=yes, ai-train=no, search=yes, use=reference: search is welcome, so is grounding an answer in these pages by reference with a link back to the page that names the person, and training is reserved. Every fact here comes from a government publisher and is cited to it, so there is nothing to withhold and every reason to want the answers found.