- Source
- ACLED — Armed Conflict Location & Event Data Project
- Cadence
- Refreshed every 30 minutes. Events filtered to the last 30 days.
Political violence and protest events compiled by ACLED from local media, partner organisations, and open-source reports. Each marker is one geolocated event. Severity is derived from reported fatalities and ACLED event type (Battles, Violence Against Civilians, Explosions/Remote Violence, Strategic Developments).
Known limitations
- —ACLED relies on reported incidents — undercount is expected in information-dark theatres (North Korea, parts of Myanmar, rural Sahel).
- —Initial event locations can be approximate; ACLED revises coordinates as more sources corroborate.
- —Fatality counts are conservative estimates, not official casualty figures.
- Source
- ADS-B Exchange /mil/ endpoint (preferred) with fallback to OpenSky Network
- Cadence
- Refreshed every 60 seconds when the feed is healthy.
Aircraft broadcasting on ADS-B identified as military by two independent signals: (1) the ICAO-24 hex address falls inside a block allocated by ICAO to a national defence force, or (2) the callsign matches an unambiguous military token (e.g. RCH, REACH, BONE, DRACO) followed by a numeric flight identifier. Aircraft on the ground are excluded. Heading, altitude and groundspeed come directly from the ADS-B state vector.
Known limitations
- —Military aircraft routinely fly without ADS-B or with it transmitting on civilian hex codes — absence of a marker does not imply absence of aircraft.
- —Earlier builds of this filter used broad callsign prefixes (KING, EAGLE, TIGER, SAM) which overlapped civilian carriers; those tokens were removed on 2026-04-18.
- —Exercises, training sorties and VIP lift account for most visible traffic — this is flight activity, not combat activity.
- Source
- Open-source reporting, navy press releases and satellite imagery analysts
- Cadence
- Updated manually as new public deployments are confirmed.
Capital ships and notable surface combatants whose general position has been reported in open sources (US Navy / Royal Navy / Marine nationale communications, NATO exercise announcements, reputable OSINT trackers). Markers show last-known area of operation, not real-time AIS.
Known limitations
- —Positions are approximate and can lag actual movements by hours or days.
- —Naval vessels frequently operate AIS-off in contested waters; this layer does not claim continuous tracking.
- —Commercial AIS feeds (AISHub, AISStream) are not yet integrated — when that changes, this layer will be re-labelled LIVE and the cadence updated.
- Source
- Consolidated installations dataset (War Room editorial)
- Cadence
- Reviewed quarterly. Labelled STATIC in the legend.
Known permanent and semi-permanent installations — airbases, naval stations, major garrisons, training centres. Sourced from open-source country defence postures, wikipedia's Military Installations categories, and public force-posture reporting. Each marker carries operator, branch and (where known) assigned units.
Known limitations
- —Coordinates for some installations are approximate (centre-of-site or nearest town).
- —Closures, reactivations and new construction may lag the published dataset by one quarter.
- —Not every foreign deployment location is publicly disclosed; some legitimate sites are intentionally excluded.
- Source
- NASA FIRMS · VIIRS S-NPP near-real-time
- Cadence
- Updated roughly every 3 hours (satellite overpass dependent).
Thermal hotspots detected by the VIIRS sensor aboard Suomi-NPP. FIRMS detects any anomaly above ~320K — active fires, industrial heat, flare stacks, and in conflict zones, munitions impacts and large vehicle fires. The Situation Board filters FIRMS detections to within 50 km of an ACLED event from the last seven days, which suppresses the dominant signal of seasonal agricultural burning in the Sahel and East Africa.
Known limitations
- —A thermal anomaly is not a battle. Even with proximity filtering, many detections will correspond to wildfires, oil infrastructure, or controlled burns.
- —Cloud cover, sun glint and smoke plumes cause both false negatives and false positives.
- —Earlier builds displayed unfiltered FIRMS data worldwide, which over-represented agricultural activity as conflict. This was corrected on 2026-04-18.