Evidence & Method

SITREP aggregates public reporting and uses automated classification. It is not an independently verified account of every incident.

Publication labels

reported

A provisional source report with a coarse location. Independent confirmation is not established.

verified

The record passed the pipeline's incident/location gate. This label is not a guarantee of independent corroboration or factual certainty.

aggregate_verified

A broad-area or multi-location report passed the aggregate location gate. It is not a precise single-strike coordinate.

needs_review

Insufficient or ambiguous evidence; not authorized as a located public map event.

Sources and corroboration

Report sources are publishers retained with the individual record. Same-city/day source diversity is a separate context measure and does not corroborate an incident. Automated matching uses time, event terms and specific-subject overlap; it can make mistakes and does not prove publisher independence. Copied reporting and source dependency can inflate apparent support. Confidence labels are pipeline estimates, not probabilities.

Location, dates and deduplication

Map coordinates are deliberately coarse, city/region-level approximations, not targeting-grade data. Multiple reports may share a marker. Automated deduplication is conservative; ambiguous pairs stay separate, and residual duplicates are expected. Publication time and collection time are different; neither necessarily establishes the incident time.

Coverage and limitations

RSS and Telegram coverage is incomplete. Absence from the map is not evidence that an incident did not occur. AI classification, translation, casualty extraction and matching may be wrong. Review the linked original reports before relying on critical claims. The curated directory is not identical to the observed source inventory. Independent/exiled Russian media are distinguished from state or pro-Kremlin outlets.