Provisional data

Provisional data is information released before the source considers its review or reporting process complete. It can provide an earlier view of activity, but values may change as records arrive or errors are resolved. The provisional label should explain what remains unfinished and which parts of the release may be revised.

Early information has a stated status

The U.S. Geological Survey's provisional-data statement explains that its observations remain subject to revision until review and final approval. In a different setting, CDC describes how reporting delays can leave recent public-health observations incomplete. These examples illustrate why an early release needs context; they do not set a universal rule for public-safety datasets. USGS statement, CDC reporting-delay explanation.

A police, fire, or EMS publisher should identify its own process. “Provisional” might concern pending validation, late submissions, or incomplete coverage. Readers should not assume every agency uses the label in the same way.

Example

Hypothetical: A department publishes a preliminary monthly count of 240 incidents. After receiving late records and correcting a duplicate, it releases a revised count of 248.

The first figure described the information available at the earlier release. The second reflects additional processing. Comparing the provisional 240 with a fully reviewed count from another month could create a misleading impression if the difference in completeness is ignored.

Common confusion

Provisional does not mean useless or wholly unverified. Some checks may already have occurred. The label means the source has not completed the process described in its documentation.

It also does not mean estimated. A provisional release can contain directly recorded observations that remain incomplete. A modeled estimate can be based on either provisional or finalized inputs. The status of the observations and the use of estimation are separate questions.

Finalized data is not a promise that no later correction is possible. It means the specified finalization process has been completed, subject to the source's policy.

Comparing early releases

Look for the reporting cutoff, included agencies, expected revision cycle, and known gaps. Preserve the release date when quoting a number. If available, compare figures at a similar stage of reporting rather than treating a very recent count as equally complete as a much older one.

A decline at the newest end of a chart deserves particular attention to reporting status. Before describing it as a change in activity, check whether records for that period are still expected to arrive.

Related terms

• Revised data
• Reporting lag
• Data validation
• Data completeness

Sources

• U.S. Geological Survey: Provisional Data Statement — Example of a source distinguishing early observations from reviewed, approved data.
• CDC: Behind the Model—Nowcasting, January 30, 2026 — Explains how delayed reporting affects recent observations; no public-safety finalization deadline is inferred.