Food Inspections — Chicago: full dataset

This information is derived from inspections of restaurants and other food establishments in Chicago from January 1, 2010 to the present. Inspections are performed by staff from the Chicago Department of Public Health’s Food Protection Program using a standardized procedure. This Actor returns it as one flat JSON row per record, with 22 declared fields — typed, named and documented in the dataset schema, so a spreadsheet or an agent can use it without guessing.

What you get

Field	Type	Meaning
inspection_date	string	Inspection Date
inspection_type	string	Inspection Type
location_address	string	Location (address)
inspection_id	number	Inspection ID
license_	number	License #
dba_name	string	DBA Name
violations	string	Violations
longitude	number	Longitude
latitude	number	Latitude
results	string	Results
risk	string	Risk
latitude_2	number	Latitude
longitude_2	number	Longitude
state	string	State
location_city	string	Location (city)
facility_type	string	Facility Type
city	string	City
aka_name	string	AKA Name
location_zip	string	Location (zip)
location_state	string	Location (state)

The output schema lists all 22 fields; the table above is the first twenty.

Every row also carries sourceRowId, sourceUpdatedAt and scrapedAt, so a pipeline can tell one record from another and know exactly when it was read.

{
  "inspection_date": "2026-10-23T00:00:00.000",
  "inspection_type": "License",
  "location_address": "text",
  "inspection_id": 2645731,
  "license_": 3103962,
  "dba_name": "BILLAL'S",
  "violations": "47. FOOD & NON-FOOD CONTACT SURFACES CLEANABLE, PROPERLY DESIGNED, CONSTRUCTED & USED - Comments: OBSERVED UNTREATED WOO",
  "longitude": -87.64111966683218
}

Inputs

Field	What it does
Filter (SoQL)	Applied at the source, so a narrow question fetches a narrow page instead of downloading everything.
Order	Empty keeps the publisher's own order; set a column to sort, for example amount DESC.
Maximum rows	Hard cap. The source is paged 1,000 rows at a time.
Time budget	Stops cleanly and reports how far it got.
Pause between pages	Keeps the run inside the publisher's rate limits.

Limitations

• This is one dataset, from one publisher. It is the Food Inspections dataset as City of Chicago publishes it — not a national or cross-city equivalent, and not a merge with other sources.
• The publisher can change it. Columns are added, renamed and retired upstream; the field list here is the schema the Actor was generated from. A renamed column arrives as a new field only after the Actor is updated.
• Fields are typed from live records. A column that was empty in the sample is declared as text; a value that does not fit the declared type comes back null rather than being coerced into a wrong number.
• Records are what the publisher publishes. No geocoding, no de-duplication across datasets, no enrichment.
• Freshness is the publisher's. The data was last changed upstream on 2026-10-25. Where the publisher supplies a row timestamp it appears as sourceUpdatedAt; scrapedAt records when this Actor read the row.

Why this is different

• Food Inspections — Chicago: full dataset declares all 22 of its output fields in the dataset schema, each with a type, a description and an example.
• The Actor reads the Food Inspections dataset directly from Chicago Open Data; no key, login or account is needed.
• Filters run at the source with SoQL, so a narrow request returns a narrow page rather than a full download.
• Every row is flat JSON: nested location objects become latitude and longitude, and no field holds a nested object.
• The row id is stable, so the same request de-duplicates instead of returning the same record twice.

FAQ

Do I need an API key? No. Chicago Open Data publishes this dataset openly.

How fresh is it? The Actor reads the dataset live. scrapedAt records when each row was read, and sourceUpdatedAt carries the publisher's own row timestamp where it supplies one.

Can I filter by date or status? Yes — use the Filter field with SoQL, for example status = 'Active'.

What happens if my filter is wrong? The run stops with the portal's own message and the hint to check SoQL, rather than returning an empty dataset that looks like "no data".

Is this the whole dataset? Up to your row cap, paged 1,000 at a time. Raise the cap to read more.

What is sourceRowId? The publisher's own row id when it supplies one, otherwise a stable hash of the row. Either way it is what makes de-duplication and joins possible.

Pricing

Pay per result: $0.0005 per row. Each stored dataset row is billed, including any diagnostic row returned instead of results.

An Actor-start event is also charged at $0.0001 per GB of allocated memory (minimum one event), even if no rows are returned. At 1 GB or less, 1,000 result rows cost $0.5001 before any applicable Store discount.

Data source

Chicago Open Data — Food Inspections, published by City of Chicago. Public data under the portal's open-data terms.