DataSF and San Francisco's Open Data Law
San Francisco's portal opened in 2009 under Mayor Newsom's Executive Directive 09-06. A 2010 ordinance put the duty into Administrative Code Chapter 22D; a 2013 ordinance added a Chief Data Officer and department data coordinators. Unusually for a US city, the data carries a recognised open licence, PDDL 1.0, and the inventory Chapter 22D requires is itself published as a dataset.
Live
📅 since 2009
✔ Editor-reviewed
📊 Impact
659 datasets in the portal catalogue, read 26 July 2026; police incident reports are republished daily and were current that day. The inventory the ordinance requires is public and was refreshed the same day: of 1,308 inventoried datasets, 696 are marked published, 474 not published, 77 internal, 51 in progress and 7 in scoping. 153 published datasets that are not flagged historical had gone more than a year without a data update.
🎓 Lesson
Chicago's duty came by executive order and the compliance reports stopped after two years. San Francisco put the same duty in the code and publishes the compliance record itself as a dataset — which is why anyone can count the 474 inventoried datasets that are still unpublished without asking the city.
Evidence🏛️ Official
Scale🏙️ City-wide
Sources
↗ Sunlight Open Data Policy Hub — San Francisco, CA, Executive Directive 09-06 (21 October 2009)The origin instrument, independently transcribed and analysed. Date: 'October 21, 2009', issued by Mayor Gavin Newsom. Carrying duties: 'All Department Heads or their designees shall strive to publish all datasets under their authority consistent with the DataSF authorization policy', with quarterly reviews of progress, and 'All datasets determined to be accessible to the public now and in the future shall be made available through DataSF.org.' The directive names the portal but does not date its launch.
↗ City and County of San Francisco — open data legislation text (Chapter 22D), city GitHub repositoryThe operative text, published by the city itself. SEC. 22D.2(a) makes the Chief Data Officer responsible for coordinating the portal, assisting departments with compliance and analysing usage, and (a)(5) to 'Present an annual updated citywide implementation plan to COIT, the Mayor, and…' the Board. SEC. 22D.2(b)(3) requires each department Data Coordinator to 'Prepare an Open Data plan for the Department', including 'a summary description of all data sets under the control of each…' department and 'a timeline for the publication of the Department's open data'. This is the sentence-level basis for the inventory and publishing-plan machinery.
↗ Sunlight Open Data Policy Hub — San Francisco, CA, 2013 revision to Chapter 22D (28 March 2013)The 2013 strengthening. Ordinance numbers reported in secondary summaries of this chain — Ordinance 293-10 (2010) codifying the policy as Administrative Code Chapter 22D, Ordinance 69-13 passed 16 April 2013 creating the Chief Data Officer and Department Data Coordinators — were not confirmed against the clerk's file, which is why the entry body says 'a 2010 ordinance' and 'a 2013 ordinance' rather than citing numbers.
↗ American Legal Publishing — San Francisco Administrative Code, Chapter 22D: Open Data PolicyThe codified chapter as it stands today, including SEC. 22D.2 (Chief Data Officer and City Departments) and SEC. 22D.3 (rules and technical standards). Blocks automated fetching (403); text verified through the city's own published legislation file above.
↗ DataSF — Terms of UseThe licence grant, and the reason this entry is distinctive among the US cities in the atlas. Carrying sentence: 'Except where otherwise stated in the file containing such Data or on the page from which such Data is accessed, including its metadata, Data is made available under the Public Domain Dedication and License v1.0 whose full text can be found at: http://www.opendatacommons.org/licenses/pddl/1.0/'. The terms also bar users from attempting to re-identify individuals in the data. sf.gov returns an empty body to automated fetching; text verified through indexed search and corroborated at field level by the dataset metadata below.
↗ DataSF — dataset metadata API, Police Department Incident Reports: 2018 to PresentPrimary licence and currency check, read 26 July 2026. Licence at field level: licenseId 'PDDL', licenseName 'Open Data Commons Public Domain Dedication and License', termsLink 'http://opendatacommons.org/licenses/pddl/1.0/' — an actual open deed on the asset, not a terms-of-use pointer. Currency: rowsUpdatedAt 2026-07-26T14:48:45Z, 1,049,895 records through 25 July 2026, metadata stating data change frequency hourly and publishing frequency daily.
↗ Socrata Discovery API — DataSF catalogueScale figure, read 26 July 2026: resultSetSize 659 datasets. Counted with the same only=dataset filter used for the Chicago entry, so the two figures are directly comparable; the filter excludes charts, maps, filtered views and files. Every result sampled carried the PDDL licence string.
↗ DataSF — Dataset inventory (y8fp-fbf5)The accountability artefact, and what San Francisco publishes that the peer cities do not: the Chapter 22D inventory is itself an open dataset, with one row per departmental dataset carrying publishing_status, department, priority, date_published, update_frequency and days_since_data_update. Read 26 July 2026: data_updated_at 2026-07-26T02:12:31Z. The dataset's own description states it is 'collected in accordance with Chapter 22D of the Administrative Code'.
↗ DataSF inventory — publishing status counts (SODA aggregate query)Sentence-level source for every number in the impact field, read 26 July 2026: Published 696, Not Published 474, Internal 77, In Progress 51, Scoping 7, blank 3 — 1,308 records. The 696 'published' inventory records exceed the 659 live catalogue datasets; the two count different things (departmental publishing records versus current portal assets) and are reported separately rather than reconciled. The staleness figure comes from a second query on the same dataset: publishing_status='Published' plus days_since_data_update over 365 and update_frequency not historical returns 153.
↗ Open Data Commons — Public Domain Dedication and License (PDDL) v1.0What the licence actually grants, for readers comparing the atlas's licence positions: a dedication of the database to the public domain, with no attribution and no share-alike condition. This is the most permissive position of any city in the atlas — more permissive than Vienna's CC BY 4.0, and unlike New York (statutory ban on licence conditions but no deed shipped), Chicago (no deed, plus a notice condition on reusers), Jakarta (asserted public-domain status, no deed), Shanghai and Bangkok (no licence at all).
↗ City Administrator announces new leadership for San Francisco data and technology initiativesThe Chief Data Officer post exists and is filled: Soumya Kalra was named Chief Data Officer and began on 13 January 2025, overseeing DataSF's open data platform and citywide data standards. Directly relevant because Chicago's Inspector General found the equivalent post vacant for 18 months. sf.gov returns an empty body to automated fetching; verified through two independent search retrievals of this release.
↗ Committee on Information Technology (COIT) — presentation, 17 April 2025Currency, and a sign the portal is still the city's compliance vehicle rather than a legacy site: the presentation sets 18 July 2025 as the deadline for the CIO's initial publication of the AI inventory on the City's Open Data Portal, with a comprehensive inventory update due 18 January 2026. Kept as context; the entry makes no claim about whether those deadlines were met.
↗ Eviction in San Francisco — Parker ZieglerExample of downstream reuse, labelled honestly as an example and deliberately kept out of the impact field. The page credits its data to SF OpenData and the US Census Bureau; the Anti-Eviction Mapping Project, active since 2013, works from the same Rent Board eviction-notice records published on the portal. The crediting sentence was read from the search index rather than a successful fetch. No source ranks the portal's usage, so no 'most used' claim is made.
📎 Cite this project
1001 Smart Cities (2026). “DataSF and San Francisco's Open Data Law” — San Francisco, United States. The Smart City Atlas. https://1001smartcities.org/projects/sf-datasf/ (last verified 2026-07-26). Data: CC BY 4.0.
The underlying data is free to reuse with attribution — see the open data page.
🔗 Related projects
🇺🇸 Chicago's Open Data Executive Order and Portal · Chicago
The same duty, two instruments: Chicago's executive order stopped reporting in 2014; San Francisco's code publishes the inventory itself.
🇺🇸 NYC Open Data and the 2012 Open Data Law · New York City
Two ways to be open: New York's law bans licence conditions and ships no deed, San Francisco ships PDDL 1.0 on the dataset metadata.
🇺🇸 NextMuni: Real-Time Muni Predictions & the Fight Over Who Owned Them · San Francisco
The exception in the same city: transit predictions arrived only after a licensing fight, and SFMTA's own feed deed stays revocable.
🇹🇭 Open Bangkok: Data Catalogue & Policy Scorecard · Bangkok
Opposite ends of the licence scale: Bangkok's files are downloadable with no licence at all, San Francisco's carry a public-domain deed.
→ Open this project in the interactive atlas · → 📡 Data & IoT category · 🚩 Report an error