
Guideline 19 asks governments to let anyone download an entire dataset at once, not just search results or filtered slices. Bulk access is the simplest, most reliable way to support full reuse and analysis.
Bulk data is unglamorous. It is a complete file, offered for download, with nothing between the user and the records. That simplicity is exactly why open data advocates treat it as the foundation everything else is built on.
What the guideline means
A search box returns what the publisher anticipated someone would ask for. A bulk download returns everything, and lets the user ask questions the publisher never imagined. Guideline 19 asks open data policies to require that public datasets be downloadable in full, whether through a portal feature or a plain "download" link on an agency page.
Bulk access is distinct from, and comes before, APIs (see Guideline 20). An API is excellent for live lookups; a full file is what analysts, researchers and model builders need to compute totals, find patterns and join datasets together.
Model policy language patterns
Bulk provisions are usually short. Paraphrased forms include:
- A definition of open data that lists "available in bulk" alongside machine-readable and open format.
- A directive that published datasets be available for complete download.
- A requirement that public data be retrievable, downloadable in bulk and reusable with commonly available tools.
Some policies fold bulk access into the duties of an oversight committee, listing it among the standards departments must meet.
Where it appears in adopted policies
In this index, bulk-data language appears in:
- Charlotte, NC (2015), whose IT policy defined open data as machine readable, available in bulk and provided in an open format such as CSV.
- Fort Collins, CO (2017), which committed to data that can be downloaded in bulk and reused with readily available software.
- Jackson, MS (2015) and Las Vegas, NV (2016), both of which tasked their governance bodies with ensuring datasets are downloadable.
- Milwaukee, WI (2016), Minneapolis, MN (2014) and Naperville, IL (2016).
Common pitfalls
- Row limits. Portals that cap exports at a few thousand rows fail the guideline in practice.
- Search-only interfaces. A web form that returns one record at a time is not bulk access, however polished it looks.
- Undocumented snapshots. A full file without a date stamp or change notes makes it hard to tell whether the download is current.
- Format lock-in. Bulk files should be in open formats (see Guideline 8), not proprietary exports.
A practical checklist for bulk access
- Define open data to include availability in bulk, not just machine readability.
- Prohibit export caps that prevent downloading the full dataset.
- Date-stamp every bulk file and note what changed since the last release.
- For very large datasets, offer dated partitions plus a documented way to obtain the complete set.
- Use open, widely supported formats such as CSV or JSON alongside any specialized formats.
- Pair bulk publication with review for sensitive fields, since a full release cannot be partially recalled.
Why this matters for your business data
Bulk data is also how operational data creates value in the private sector. AI and machine-learning teams rarely want a dashboard or a query endpoint; they want a complete, consistent extract covering a long period, with a schema they can rely on. A business that can produce a clean, full export of its work orders, transactions or logs, de-identified and documented, has an asset that buyers can actually evaluate.
The questions are the same ones a city faces: is the export complete, is it in an open format, is it dated, and is anything sensitive removed? The de-identified data guide covers the last point, and get a bid is the place to start if you want a buyer's view of what your extract is worth.
We never take HIPAA/PHI data or personal customer data. Everything is de-identified before it leaves your control.
Frequently asked questions
What counts as bulk data?
Bulk data means the entire dataset can be downloaded in one action, in a machine-readable format, without row limits or one-record-at-a-time interfaces. Large datasets may be split into dated files, as long as the full set is available and documented.
Should governments offer APIs instead of bulk downloads?
The guidelines treat the two as complementary, with bulk access as the baseline. APIs serve live, targeted queries; bulk files serve full analysis. An API alone forces users to reconstruct the dataset through many requests, which is slow and error-prone.
Are there privacy concerns with bulk releases?
Yes, which is why bulk publication must follow review for sensitive information. Once a dataset is released in full, mistakes cannot be recalled. Policies typically pair bulk-access requirements with safeguards like those described in Guideline 7.
Policies in the index that reflect guideline 19
| Place | Adopted | Instrument | Guidelines |
|---|---|---|---|
| Minneapolis, MN | Aug 1, 2014 | Council-adopted policy | 34568131819 |
| Charlotte, NC | Jan 1, 2015 | IT policy | 2781924 |
| Jackson, MS | Sep 1, 2015 | Executive order (No. 2015-06) | 12345678 |
| Durham, NC | Nov 23, 2015 | Administrative policy | 7818192430 |
| Las Vegas, NV | Apr 12, 2016 | Administrative policy (City Manager policy CM510) | 12345678 |
| Milwaukee, WI | Jul 14, 2016 | Council resolution | 12356789 |
| Naperville, IL | Nov 15, 2016 | Council resolution | 45789101112 |
| Fort Collins, CO | Feb 7, 2017 | Council resolution | 12345678 |
| Knoxville, TN | May 2, 2017 | Executive order | 14578111319 |
Sources
Updated December 31, 2024.