Contact

GlossaryPublic records, privacy, and data quality

Metadata

Metadata explains a dataset's source, period, fields, and processing. Learn how it helps readers interpret a public report and distinguish different update dates.

Metadata is information that describes data. It can identify who produced a dataset, what it contains, when and where it applies, and how it was collected or processed. For a public-safety report, metadata provides the context needed to understand what a number or record actually represents.

The explanation around the rows

The U.S. Geological Survey describes metadata as documentation of the who, what, where, when, why, and processing behind data, including quality controls. It can be provided in forms readable by people and by software. USGS: What are metadata?.

On a public portal, metadata may appear as a description, field dictionary, source contact, date range, license, or methodology link. These details are part of interpreting the file, even when they sit outside the downloadable rows.

Example

Hypothetical: A file named “Annual Incidents” contains 8,000 rows. Its metadata says each row is an offense record, the file covers one department, and the dates refer to when reports were filed.

Those details change the reading. The 8,000 rows are not automatically 8,000 separate incidents, the coverage is not automatically countywide, and the annual period does not necessarily identify when every event occurred. The filename alone would leave all three questions unresolved.

Common confusion

Metadata is broader than a data dictionary. The dictionary usually explains fields and values; other metadata describes the dataset's origin, coverage, release, and use conditions.

Metadata also has its own update history. A portal can revise a description or contact information without changing the underlying records. A “metadata updated” timestamp should not be read as proof that new incidents were added.

Documentation is not an automatic quality guarantee. It can describe known limitations accurately, or it can be incomplete or outdated. Compare the stated definitions with the file and investigate inconsistencies before relying on the result.

What to keep with an analysis

Preserve the dataset title, source organization, version or download date, field definitions, and relevant exclusions. If a calculation depends on a particular date field or counting unit, carry that explanation into the published chart.

This is especially useful when a dataset changes. A later reader may otherwise see different fields or revised records at the same URL and be unable to reproduce the earlier result. Good metadata keeps the connection between the data and its meaning visible across downloads, summaries, and future revisions.

Sources