Methodology & Data Sources

What is the Global Armed Conflicts Map?

The Global Armed Conflicts Map is a free, open intelligence dashboard that tracks 29 active armed conflict theaters worldwide on a single interactive map, with a plain-language intelligence briefing for each. It exists to bridge the gap between dense academic conflict datasets and digestible, at-a-glance situational awareness.

Every theater is profiled with the same structure: key actors, conflict type, current status, intensity classification, geography, what is at stake, a timeline, and source attribution. This consistency makes conflicts comparable side by side.

Where does the data come from?

We aggregate four primary, widely cited sources rather than collecting raw events ourselves:

• ACLED (Armed Conflict Location & Event Data Project) — event-based, near real-time geolocated records of political violence. The unit is the individual event; there is no fatality threshold.
• UCDP (Uppsala Conflict Data Program) — conflict-centric and historical. The unit is the conflict dyad, with a threshold of ≥25 battle-related deaths per year and annual updates.
• CFR Global Conflict Tracker (Council on Foreign Relations) — strategic, qualitative briefings on status and impact.
• Google News — aggregated headlines from global news outlets, which feed the automatically updated "Latest developments" on each theater page.

For a full side-by-side comparison of how the three core datasets differ in unit, threshold, update cadence and intended use, see ACLED vs UCDP vs CFR.

How we organize conflicts: the "theater" unit

The dashboard's organizing unit is the conflict theater — a coherent war or crisis (such as the Russia–Ukraine war or the Sahel insurgency) — rather than the individual event (ACLED) or the academic dyad (UCDP). This is a deliberate editorial choice: it is how journalists, analysts and the public actually think about wars, and it lets a transnational conflict like the Sahel insurgency be presented as one theater instead of three national datasets.

How are status and intensity classified?

Status (Active, Tensions, Frozen) and intensity (High, Medium, Low/Elevated) are display classifications, not source-reported fields. They are editorial summaries based on the underlying sources, intended for orientation. A standoff such as the Taiwan Strait is labelled "Tensions" precisely because battle-death datasets record little there — open war has not occurred — even though strategic risk is high.

How often is it updated?

The "Latest developments" section on each theater page is refreshed automatically every few hours from Google News, and each page's dateModified updates with it. The structured profiles (actors, timelines, stakes) are evergreen and reviewed periodically. Casualty figures everywhere on the site are estimates drawn from contested reporting, not confirmed counts.

The build itself runs every eight hours. Each run refreshes the headlines, checks UCDP for a newer month of reported-deaths data (see Reported deaths below), regenerates every page, and then runs three automated checks for broken URLs, stale years and search-metadata regressions before the result is published.

Reported deaths (UCDP)

The homepage globe shows a reported-deaths chart for each theater: a twelve-month total, the share of it recorded as civilian deaths, and a monthly bar chart. Those numbers come from one source only, the UCDP Candidate Events Dataset, which the Uppsala Conflict Data Program publishes as one open file per month, roughly a month after the month ends. We use UCDP's best estimate of fatalities for each event and sum events by the month they started. Where UCDP records a range, the low and high ends are kept as well, and deaths UCDP could not attribute to a side or to civilians are counted as "unknown" rather than dropped.

Each theater is mapped to one or more UCDP country codes, and that mapping is deliberately simple and public (it is the table at the top of scripts/fetch_fatalities.py in the site's repository). A transnational theater such as the Sahel insurgency sums Mali, Burkina Faso and Niger. Where two theaters share a country, the country goes to one of them: UCDP codes events in Gaza and the West Bank under Israel, so Israel's events count toward the Israel–Gaza war and not the Iran–Israel confrontation; the Kashmir theater counts India only, because Pakistan is its own theater. A maritime standoff such as the South China Sea has no country mapping and shows none.

• Preliminary by design. Candidate data is UCDP's first pass. It is revised, sometimes substantially, in the annual release each June, so a month's total can change after we show it.
• Direct deaths only. UCDP counts deaths from organized violence. Hunger, disease and displacement are not in these numbers, and in Sudan or Gaza they may exceed the direct toll.
• A floor, not a total. Where access is blocked, UCDP records what can be sourced. A small number on the chart can mean little violence or little reporting; the "unknown" share and the low–high bracket are the honest signals of which.
• About a month behind. The chart states the last month it covers. The build checks for a newer month every eight hours and swaps it in when UCDP publishes.

Atrocity determinations

Each theater panel on the globe also carries a four-step atrocity indicator, and the sidebar marks some theaters "Genocide finding" or "Genocide case". This is not a judgement by this site and it is not derived from death counts. It records the most serious kind of formal finding that a court, a United Nations body or a government has made about the conflict, and it cites the authority and the date of each finding so it can be checked.

Three rules govern the indicator. The level describes the conflict's current phase: a determination about an earlier phase, such as the genocide findings against ISIS for 2014–17 in Iraq and Syria, is kept as a cited finding but does not set today's level. A level only rises with a citable authority and date; opinion, advocacy and reporting do not count, however credible. And where the party found against rejects the finding, the panel says so. The full list, with every citation, is the file data/atrocity_determinations.json in the site's repository, and corrections are welcome through the corrections page.

Limitations and caveats

• Estimates, not ground truth. Fatality and displacement figures are approximate and often contested between sources.
• Reporting gaps. Conflicts such as Sudan and Myanmar are systematically underreported due to access and communications blackouts, so figures likely understate the true toll.
• Secondary aggregation. We are not a primary data collector; always cite the underlying sources (ACLED, UCDP, CFR) for high-stakes work.
• Editorial classifications. Status and intensity labels are our summaries, not official designations.
• The globe's death counts are one dataset. They are UCDP candidate data, preliminary and direct-deaths-only, and they will differ from the estimates quoted in the briefings, which draw on several sources. The atrocity indicator records formal findings, not the scale of suffering: a level 0 theater can be deadlier than a level 4 one.

How to cite this dashboard

When referencing figures, please attribute the underlying sources directly. To cite the dashboard itself: "Global Armed Conflicts Map, armedconflicts.org (accessed [date])." We welcome use by journalists, researchers, educators and analysts — see who the map is for.

Frequently asked questions