chiwardboard.com

About

ChiWardBoard ranks Chicago’s 50 wards by how fast the city closes 311 service requests filed in them - towing an abandoned car, filling a pothole, baiting an alley. The records are public. The comparison was not. Now it is.

The numbers come from the City of Chicago’s 311 Service Requests dataset. Nothing here is our own measurement. Anyone can check the figures against the source.

A ward that closes requests quickly is not necessarily a better-run ward. Some of these gaps come from how much arterial street a ward has, or how many people file requests in the first place. The methodology below says what these numbers can and cannot tell you.

Every ward page carries that ward office’s published contact information, along with a link to file a 311 request yourself, because the numbers only help if you know who to call.

Not affiliated with the City of Chicago. Free and open to anyone, with no account and no ads. Corrections welcome at after.hours.builder@gmail.com.

Methodology

• Period. A rolling 12 months ending at the last complete month, so a part-filled month never distorts a figure; the city’s own published median dataset uses the same shape. 2024 and 2025 can also be selected. Earlier years are not offered: the ward map changed in May 2023, and comparing across it would compare different areas under the same ward numbers.
• Who represents a ward. The alderperson and contact details on each ward page come from the city’s Ward Offices directory. Links that do not work - a dead domain, an email miskeyed into the website field - are corrected or dropped, and the build warns when another stops resolving; everything else is published as the city has it. The period can span more than one officeholder, so a ward’s figures are not necessarily the work of the person named.
• Source. The City of Chicago’s public 311 Service Requests dataset and its 2023 ward boundaries, plus the traffic crash and bike route datasets for the cycling context. Everything is computed from the rows at build time and published as a dated snapshot, readable here; every ward page links to the live rows behind its figures.
• “Closed” means Completed. Cancelled and still-open requests are not thrown away - each counts as a wait of at least that long. Dropping them would make a ward’s unfinished work vanish from its own figures and look like speed.
• Duplicates are excluded. A duplicate is the same physical problem reported twice, so counting it would time one repair as two. The city excludes them in its own tooling. The number excluded is shown with each board.
• Days to close. Closed timestamp minus created timestamp. For a request still open, the wait so far, counted as at least that long; a cancelled one the same, up to its close. The typical time is estimated from all of them together, so the unfinished ones pull it where they belong. Negative durations and rows with no ward are dropped, with the counts printed beside the claim they affect.
• Two ranking floors. A ward is ranked on a type once it has 100 completed requests of it, and only where enough closed for a typical time to exist; below that it is listed and flagged rather than ranked. The fastest and slowest named at the top of a board answer to a higher bar of 200. The difference is deliberate: an endpoint is chosen by taking a maximum, which is the one operation guaranteed to surface a fluke, and on the thinnest types that turned a 0.05-day margin into a claim about the fastest ward in Chicago. A ward between the two floors is ranked in the table and never named at the top.
• One board is measured the other way round. Almost every type here closes eventually and differs only in speed: over two years, the share of potholes, rat complaints or tree debris still open after six months is effectively zero. Sidewalk repairs are the exception in the whole dataset - the city completes too few to rank on speed, and about a third sit unfinished. So that board ranks the share still open, over a 30-month window, counting only requests filed at least 180 days ago, because a request from last month is new rather than late. The window is long because that 180-day wait eats into it; over a single year a typical ward has too few requests left to measure a share against. Wards with fewer than 100 old enough to judge are shown but not ranked. Same question, answer inverted, labelled as such wherever it appears.
• What an open sidewalk request proves. Closing one is not a formality: since the May 2023 ward map the city has completed 19,691 of them, taking a median of 165 days, with one in six hundred closed the same day. So an unfinished one is worth counting. But the route to a repair matters too - for a residential sidewalk it usually runs through CDOT’s Shared Cost Sidewalk Program, which is voluntary, first-come first-served, capped by funding, and opens for applications one day each January. We publish an open request as work the city has not signed off on, and say no more than that.
• Individual requests, and only on city property. Open more than a year lists specific requests with the address and case number, oldest first, nothing chosen for being interesting. Every type on it is a complaint about something the city owns, and has to be work the city is normally slow at - it closes a traffic signal complaint in a median of five hours, so one open for a year says more about the paperwork than the intersection. Types like that are measured out on every refresh rather than listed by hand. Complaints against private property and its owners are never listed. A request can also be shown as having sat through several checks months apart; that count starts when the watching started.
• Cycling figures are context, not a ranking. Ward pages carry crashes involving someone on a bike, how many left a person seriously hurt or killed, and miles of bike route - placed into wards by coordinate. They are never ranked: a ward with more crashes is usually a ward with more cycling, and correcting for that needs ridership figures nobody publishes. Reported bike lane obstructions are not here at all, at about 630 a year citywide.
• Gaps over ratios. Ward differences are stated in absolute days. When the fastest ward is near zero, a ratio divides by almost nothing and becomes meaninglessly large, so we do not print one.
• What’s off the board. Informational calls and aircraft noise complaints are logged at city facilities (the 311 call center in Ward 28; O’Hare in Ward 41), so they describe a building, not a neighborhood. Types that close almost instantly stay on the board with that fact stated.

Questions & comments

Corrections, ideas, and questions: after.hours.builder@gmail.com.

Privacy

• No accounts and no ads, and the site sets no cookies at all.
• “Use my location” is resolved against the ward map inside your browser. Your coordinates are never transmitted anywhere.
• Address lookup works the same way. The site downloads a block-to-ward index built from public records, and the match happens on your device. The address you type is never sent to us or to anyone else.
• The page loads Vercel Web Analytics, which counts visits without cookies and cannot recognise you across days or across other sites. Our hosting provider (Vercel) and Google Fonts also receive standard web requests (IP address, user agent) to serve the page, as with any website.

Terms

• ChiWardBoard is an independent project and is not affiliated with the City of Chicago.
• The numbers describe how requests were recorded, which is not always how work happened. Provided as-is, for information, with no warranty. Every figure links to the rows behind it, so check them before quoting one.
• Underlying records are public data provided by the City of Chicago under its data terms.