How we collect, check and cite public records
Every dataset passes the same six steps before a figure can appear in a report, and a load that does not match the publisher’s own record count is blocked rather than published. We claim no proprietary data: these are official public records, we keep every copy we fetch, and where we estimate something we show the evidence the estimate was tested against.
Every dataset we use · where each report works · corrections we have made · free checks
Every load passes the same six steps
Acquire
Official APIs only. We confirm every field we rely on still exists, and ask the publisher how many records to expect.
Archive
Every page is saved, dated and fingerprinted before it’s read, so we can show what a publisher said on any day.
Stage
Values are converted strictly. Anything that doesn’t convert exactly is rejected with a reason, never guessed.
Validate
Record counts must match the publisher’s. Logic checks, such as building plus land equalling the total, block bad loads.
Publish
All at once or not at all. If published rows don’t match checked rows, the change rolls back.
Audit
After a full load, random published records are re-read from the publisher’s live API and compared field by field. A daily scheduled refresh reloads each source when its publisher’s update cycle is due.
Identifiers first, names last
When we match a record to your property, facility or contractor, we use official identifiers (a parcel number, a license number) before addresses, and addresses before names. Every match stores how it was made and how confident it is. Matches below our threshold go to a person for review instead of into a report.
Four states, never merged
- Sourced · value taken from a named dataset and record
- Not published by source · the publisher doesn’t record this
- Could not be read · the record exists but a value failed strict checks
- Not checked · outside what this report covers
How the Assessment Check rates a home, and the evidence behind it
We compare your assessed value per square foot of building area with the median of comparable homes: same township, neighborhood and Assessor class, building area within 20% of yours, built within 15 years. At least 5 comparable homes are required, or no result is given. Parcels with more than one building aren’t rated yet; condominium units use the separate method below.
| Result | Your value per sq ft vs comparable median | Assessor appeals granted, 2025 | Assessor appeals granted, 2024 | Board of Review appeals granted, 2024 |
|---|---|---|---|---|
| No case found | At or below the median | 9.0% of 90,515 | 9.3% of 85,296 | 33.8% of 116,822 |
| Weak | 0–5% above | 16.0% of 44,885 | 17.8% of 36,852 | 58.0% of 46,659 |
| Moderate | 5–10% above | 28.0% of 23,212 | 29.5% of 22,319 | 67.8% of 26,510 |
| Strong | 10% or more above | 40.5% of 21,997 | 40.9% of 27,562 | 71.1% of 31,695 |
Measured on 14 Sep 2026 from Cook County Assessor and Board of Review appeal records for residential homes with one building record (Assessor: 180,609 appeals in 2025 and 172,029 in 2024; Board of Review: 221,686 in 2024). “Granted” means the decided value was lower than the value appealed. The 2024 results were not used to set the levels and confirm them. Outcomes show association, not cause: reductions also follow corrected characteristics, vacancy or other evidence. Gaps of 30% or more were granted slightly less often in every test, so those results carry a prompt to check the home’s characteristics first.
Estimate. This is an estimate from public records, not a guarantee of a reduction and not legal or tax advice.
Condominium units: the building’s assessments against its own sales
Condo appeals are decided for the whole building: in 2024, every appealing unit got the same outcome in 7,722 of 7,765 buildings at the Assessor and 8,270 of 8,460 at the Board of Review. So we rate the building. For each arm’s-length sale of a unit in the 3 years before the tax year, we divide the unit’s mailed assessed value × 10 by the sale price (homes are assessed at 10% of market value). The building’s median is compared with the median of other condo buildings in the same neighborhood. At least 3 sales and 5 comparable buildings are required, or no result is given. We use the Assessor’s own flags to exclude multi-parcel sales, repeat sales within a year, sales under $10,000 and excluded deed types.
| Result | Building ratio vs neighborhood median | Assessor, 2025 | Assessor, 2024 | Board of Review, 2025 | Board of Review, 2024 |
|---|---|---|---|---|---|
| No case found | At or below the median | 3.8% of 706 | 4.4% of 1,450 | 16.4% of 996 | 27.9% of 1,578 |
| Weak | 0–5% above | 13.8% of 406 | 28.1% of 868 | 28.0% of 542 | 41.0% of 886 |
| Moderate | 5–10% above | 29.1% of 189 | 58.2% of 428 | 28.6% of 262 | 30.1% of 429 |
| Strong | 10% or more above | 32.3% of 155 | 64.4% of 292 | 30.3% of 198 | 25.1% of 283 |
Share of appealing condo buildings granted a reduction (most of their appealing units got a lower value), measured on 14 Sep 2026. The method and the rule for publishing it were written down before any appeal outcome was measured. The rule required results to rise from No case found to Strong in both Assessor years and at least one Board of Review year; that held in Assessor 2025, Assessor 2024 and Board of Review 2025. It did not hold for the Board of Review in 2024, where Weak buildings were granted reductions more often than Moderate or Strong ones. A result exists for about 40% of appealing buildings; the rest have too few recent sales or comparable buildings. An earlier condo method, comparing value per sq ft of building area, showed no link with outcomes and was not published.
Houses bought recently: the purchase price as the evidence
An owner’s own arm’s-length purchase is evidence of market value that the Cook County Board of Review asks to see (its Rule 18: the sale documents, the buyer’s and seller’s identities and any relationship, and an affidavit about the sale when the price differs substantially from the assessed value). The Tax Appeal packet uses a purchase only when every one of these holds, from the Assessor’s sales and characteristics files:
- the home is a house (Assessor class 2xx other than 299; condo units are appealed with their building);
- the sale is the most recent recorded for the PIN, on or after 1 January two years before the tax year appealed;
- the Assessor’s own sale filters pass (one parcel, not a repeat sale within a year, not under $10,000, not an excluded deed type), and the deed is a warranty or trustee’s deed;
- the Assessor class is the same as at the sale, and the building area and year built are unchanged since the sale;
- the price is 10% to 50% below the market value the mailed assessment implies (the assessed value × 10); a price further below is more often a transfer that isn’t a market sale;
- at least 3 arm’s-length sales of similar homes (same township, neighborhood and class, building area within 20%, in the two years to the latest sale in the file) back the price: it is at least 80% of their median price per square foot.
The packet then asks for the purchase price at the 10% level of assessment, shows the similar-home sales behind it, and tells the owner not to file on it if the sale was between relatives or related companies, a foreclosure, short sale or bank sale, or included other property, which the county’s records can’t show. No appeal outcomes are measured for this route yet, so the packet gives no success rate for it.
Privacy. We don’t collect owner, buyer, seller or appellant names from property datasets. We add a personal field only when a named report needs it.