The absence that shapes everything else
The most consequential fact about Chicago's published building data is something that is not in it. The City publishes no standalone Department of Buildings inspections dataset. A catalog search of the Data Portal run on October 22, 2026 for inspections and for permits returns no citywide buildings inspections table at all.
What the search does return are inspections belonging to other departments and other purposes: food inspections owned by Public Health, environmental inspections owned by the Department of Environment, a violations-and-inspections set for a targeted neighborhood recovery program rather than the city as a whole, and a life safety evaluations dataset marked deprecated in 2015. None of them is the record of whether construction work in Chicago was inspected.
The effect of that single omission is larger than every caveat attached to the datasets that do exist, and it determines what a reader can learn from an address search before they type anything.
Why a passed inspection leaves no trace
Department of Buildings inspection data reaches the public portal by one route only: where an inspection produced a violation. The inspection number, inspection status, inspection category and inspection waived fields are carried on violation rows, inside the Building Violations dataset. There is no corresponding publication path for an inspection that found nothing wrong.
So a passed inspection generally produces no public record at all. The inspection system is visible to the public almost exclusively through its failures. Every clearance the Department has ever issued on permitted work is, as far as the open data is concerned, invisible.
The clearest confirmation of this comes from the City's own watchdog. When the Office of Inspector General audited the permit inspections process, it had to obtain 249,700 inspection requests from the Department directly rather than from the portal, because the portal does not carry them. If the City's auditor could not get the data from the published record, neither can anyone else.
What that does to an address search
Two inferences readers make routinely are both unsupported by the structure of the data.
The first is that a building with no violations has a good record. It does not have a good record; it has no record. A building that was never complained about and that falls into no periodic inspection cycle produces nothing, whatever its condition. Absence of a violation entry is not evidence that an inspector found the building sound, because an inspector who finds a building sound writes nothing the public will ever see.
The second is that inspection fields on a violation row represent an inspection history. They do not. They describe the one inspection that generated that citation. Assembling them across an address gives a history of the occasions the City found fault, which is a biased sample of the occasions the City attended, which is itself a biased sample of the work that was done.
The practical upshot: the published record can tell a reader that something went wrong, and can almost never tell them that something went right.
Everything before 2006
Both principal datasets begin in 2006 — the Building Permits data covers 2006 to present, and Building Violations covers January 1, 2006 to present, both according to the City's own metadata fields. Nothing earlier is published.
For Chicago that cut is unusually severe. The housing stock is largely older than 1940, which means the public record covers the last fraction of most buildings' lives. A two-flat built in 1912 may have been converted, re-roofed, reframed, rewired and re-porched several times over before the published record starts. Those decades did generate paper. They did not generate rows in a dataset, and the City has not retrospectively digitized them into one.
This is why a short permit history on an old Chicago building is close to meaningless as a statement about the building. It is a statement about the last two decades of filings, set against a century of work that was never going to appear.
Permits that were removed from the record
The Building Permits dataset description states that it covers permits issued from 2006 to the present, excluding permits that have been voided or revoked after issuance.
That exclusion has a consequence most readers would not guess: the public record cannot show a revocation. A permit that was issued and later revoked does not appear flagged as revoked — it is absent, indistinguishable from a permit that was never applied for. The dataset cannot be used to establish that a permit was revoked, and the absence of a permit at an address cannot be used to establish that none was issued.
The enforcement machinery itself is real enough. The Municipal Code lists failure to obtain a required rough or final inspection and failure to obtain a required certificate of occupancy among the grounds for suspending a person's permit privileges under section 14A-3-304.1. What is missing is not the enforcement. It is the publication of its results.
Source: Building Permits dataset metadata, City of Chicago.
The documented pause in updates
The City publishes change notices for its datasets, and one of them matters directly to anyone reading a permit history. A notice dated September 13, 2024 is described by the City as covering a temporary pause in updates and changes to the Building Permits dataset.
That the pause occurred is documented by the City rather than inferred. How long it lasted and exactly when it began and ended could not be verified — the body text of the notice is rendered in a way that did not yield to retrieval, so only the City's own one-line description of it is available. No estimate of the length is offered here, because any would be invented.
What a reader can take from it is limited and still useful: a thin or empty stretch of permit records around that period may be an artifact of the pause rather than a reflection of what was happening at an address. Other change notices are published for the same dataset — a permit condition column added in October 2025, further changes in May 2024, and column changes in 2019 — so the schema a reader sees today is not the schema that produced every row in it.
Elsewhere on the portal, staleness is explicit rather than paused. The City's performance-metric datasets for the time taken to issue permits under the different review routes were last updated in 2014. They are evidence that the routes existed and were measured separately. They are not current performance data and should not be read as such.
What the published record can and cannot settle
Taken together, the omissions define the boundary of what any Chicago address search can establish.
It can establish that a permit was issued, on what date, under which review process, for what described scope, at what fee, in which ward and community area. It can establish that a violation was cited, under which code, on what date, and what status the file now carries.
It cannot establish that work was inspected; that it passed; that a violation was genuinely remedied rather than closed; that a permit was revoked; that no permit was issued; or that a building is currently compliant with the code. The City states expressly that its violation data does not necessarily reflect the current condition of a building or property.
That is a narrower instrument than most readers expect, and the narrowness is structural rather than a matter of searching harder. The strongest records in the Chicago system — a passed final inspection, a certificate of occupancy, which is the document the City issues when a building may lawfully be occupied — are either unpublished or confined to a class of building that excludes most Chicago homes. Knowing which questions the record cannot answer is the part that protects a reader.