To tell someone which districts contain their home, you need boundary data from 102 separate Illinois county governments. There is no statewide file, no common format, and no shared expectation that any of it belongs online. This is what that actually looks like.
districtry answers one question: who represents this spot? Click a point, get every civic district that contains it — ward, precinct, county board, fire protection district, judicial subcircuit, elected school board — and the name of the person holding each seat.
In Chicago that question is nearly easy. The city runs an open data portal, the county runs a GIS service, and the boundaries come back over HTTP in seconds. That version of the project took a weekend.
Then it left the city. Every county line the map crosses is a different publisher, a different format, a different answer to the question of whether this is public information at all. Coverage now stands at 91 of Illinois' 102 counties, and almost none of the work was code.
These are all answers to the same request — where are your voting precinct and district boundaries? — arranged by how much of the work a computer can do. Every rung is a real county's real answer. The bars on the right are the whole argument.
Machine-readable
Fetch it and you're done. Roughly a day of work per county.
A live GIS service
Cook · DuPage · Lake · Will · Kane · McHenry · Kendall
An open ArcGIS or Socrata endpoint the app queries at runtime. Boundaries stay current without anyone touching the code. This is the ceiling, and seven counties clear it.
A shapefile, on request
Jefferson · Ogle · Montgomery
Not on the website. Not in any catalogue. It arrives as an email attachment the same day somebody asks the clerk for it — proper geometry that simply had never been posted.
Human-readable
The information is public. Converting it to data is a person's afternoon — or week.
A vector PDF map
Stephenson
Two PDFs on the county's elections page — one urban, one rural — carrying real vector outlines. Thirty-six precincts were rebuilt by tracing them, then checked against the county's own board-district map, which caught a defect neither map would have shown alone.
A PDF map book
Ogle
Fifty-one pages of precinct maps, plus a polling-location dataset that is points with no boundaries. Complete, public, and useless to software.
A printed directory
Washington · Marshall
A 40-page county "Blue Book" and a five-page table of elected officers — both sent as attachments, neither findable by search, neither on a county website. The Blue Book is now the richest municipal source in the project.
Scanned images
Brown
Election summaries and precinct reports posted as pictures of pages. Readable by a person, opaque to software — not even the precinct names can be lifted, let alone the shapes.
Not available
The county holds the boundary. You cannot have it.
A licence and a price tag
Champaign · Piatt · Bureau
Champaign and Piatt sit complete and current behind the Champaign County GIS Consortium, which sells the data under a licence that permits viewing and forbids republishing. Bureau County's GIS office quoted its board-district shapefile at $150 — beneath a licence that forbids redistributing anything derived from it. The block is legal, not technical, and no amount of engineering touches it.
Refused at the door
Adams · Perry
Adams publishes its seven board districts as clean map data — and its website answers every automated visit with a 391-byte Access Denied. The map is open; the page naming who won those districts is not.
Paper — or no website at all
Edwards · Henderson
The Edwards County Clerk volunteered it unprompted: the county's voting district boundaries exist on paper. Edwards has no county website. Henderson's official address, the one in the state's own clerk directory, forwards to a domain parking screen.
After the search engines were exhausted, the remaining tool was email — one narrow question at a time, addressed to the office that would know. Clerks and GIS staff answered, often within the hour. Four exchanges that each taught the project something it had wrong.
Hamilton County Clerk's office
The question went out at 9:04 in the morning and was answered at 9:08 — mid-website-migration, with the board's structure, a GIS link, and an apology for the pages that hadn't moved over yet. Told that the McLeansboro precincts were unreadable at that scale, the office followed up with a PDF that included the polling places too.
Nothing about the county changed that morning. It had the answer the whole time. Somebody just asked.
Jefferson County Clerk's office
The precinct shapefile came back as an attachment, then the list of which precincts form each of the thirteen board districts. One question remained: what separates Dodds 1 from Dodds 2? The answer, given as the Clerk's own recollection, was that the Casey Fork river splits them — Dodds 2 west, Dodds 1 east.
Some boundaries are not in any file. No portal has a field for "the river divides them."
Montgomery County GIS
A typo spotted in the published data was confirmed and sent to IT. Then, asked which to believe when the GIS polling places disagreed with the Clerk's official list, the county's own GIS tech said the Clerk's list should take precedence — he isn't told about location changes until a new map is needed, so his layer is the older one.
The authoritative source was not the published one, and only the publisher knew. The portal said nothing.
Clark County Clerk & Recorder
Read as a refusal, that sentence ends the county. Read as information, it is half an answer — because the first half was "the County Board is elected by districts," and Clark's districts are built from whole voting precincts. The Clerk's own certified election canvasses tabulate every board race precinct by precinct, which describes each district exactly. The county went on the map four hours later, with no map involved at any point.
A county that cannot hand you the boundary may already have published everything needed to reconstruct it.
"The county publishes no precinct boundaries" is a fact about a website, not about a county.
Ogle County is the proof. Its record in this project read: the county publishes no precinct boundaries — only a 51-page PDF map book and a polling-location dataset that is points only. Every word of that was true, and had been re-verified more than once.
The shapefile arrived the day somebody asked for it. Of course it did. A county that runs elections necessarily has precinct geometry — it cannot hand out the right ballots otherwise. Putting that file on a website is a separate act, and one it may simply never have had a reason to perform.
The same logic runs through everything the map draws. A county that levies a fire protection tax knows exactly where that district ends. A county that seats a board by district knows which townships compose each one. The boundary always exists somewhere. What varies — county to county, office to office — is whether anyone ever decided it should be public.
And sometimes it is already public, in a form nobody filed under “map.” A board district built from whole voting precincts is described completely by the county’s own election returns: the list of precincts a district’s contest appeared on is the district. Four counties went on this map in a single day that way — from certified results their clerks already publish as ordinary data, with nobody answering an email. The same records were just as useful when the answer was no. Nine more counties turned out to split precincts between districts, or to have redrawn their precincts since the last census, and every one of those is now a measured fact instead of an open question.
Sometimes the honest answer really is no, and that gets recorded too. Edwards County's boundaries are on paper because that is where its Clerk says they are. That is a closed route, not an unmeasured one, and the difference matters: this project keeps a record for every absence, naming the specific artifact a publisher would have to release to close it. Currently there are 110 of them.
91
of 102 Illinois counties served
101
recorded gaps, each naming what would close it
0
counties left unresearched
Of those 106 gaps: 79 are things no one publishes in any form, 20 are published but wrong, stale, or incomplete, and 7 are blocked outright by a licence or a server that refuses to answer. Only the last seven are anybody's deliberate no.
None of this is a scandal, and no clerk in this story did anything wrong. County election offices run on small staffs and thin budgets; nobody funded a GIS portal in a county of six thousand people, and nobody should have to. The people who answered these emails were, without exception, generous with their time.
But the cost of not publishing doesn't land on the county. It lands on the resident who cannot easily find out who represents them — and it lands hardest in the smallest counties, the ones least able to build the tool that would tell them.
Which makes the fix unusually cheap. In most of these counties the file already exists, on a machine somebody signs into every day. Posting it changes nothing about the county's work and everything about what anyone can build on top of it. That is the whole ask: publish the file you already have.