LOADING

Methodology

The Myanmar Conflict Map builds on over 10,000 reports of violent incidents that have occurred in Myanmar since 1 July 2020 as collected in the Armed Conflict Location & Event Data Project (ACLED). ACLED collects data on a weekly basis from international and local media, including Burmese and ethnic media sources, then subjects it to reviews to ensure validity, accuracy, and relevance. The IISS research team has further reviewed ACLED’s data and adapted it to the Myanmar context. These adaptations account for a significant amount of uncertainty around reporting on conflict in Myanmar, particularly with regard to details such as casualties and precise locations where these incidents occurred.

Mapping incidents at the ‘township’ level

More than half of the events in the ACLED dataset are coded only to the township level or to an approximate location between village tracts, rather than to an exact latitude and longitude. The dot-density map used here takes into account the difficulty of establishing a precise location for most events in the database by allocating each event a randomised geolocation within the township in which it occurred. The map should thus be read as a representation of conflict dynamics within Myanmar and each of its states or regions, not a representation of conflict hotspots within each township.

Classifying incidents

ACLED categorises events into six types and 25 subtypes, which the research team reclassified into five event types to fit the Myanmar context:

  • attack/armed clash
  • remote explosives/improvised explosive devices
  • air/drone strike
  • crackdowns
  • infrastructure destruction

For example, the Conflict Map uses ‘crackdowns’ to encompass the ACLED event subtypes ‘protests with intervention’, ‘violent demonstration’, ‘mob violence’, and ‘excessive force against protesters’. In the Myanmar context, the terms ‘violent demonstration’ and ‘mob violence’ can be misleading, given that events in these subtypes refer to attacks by the State Administration Council (SAC) on protesters, sometimes using live rounds, and protesters’ occasional use of tear gas, barricades, and rudimentary weapons to defend themselves. The Conflict Map recodes these events as ‘crackdowns’ to reflect the asymmetry of violence between the SAC and civilians.

The Conflict Map excludes ACLED data in two instances: to remove non-violent events; and when events in a particular subtype are likely to have been significantly underreported. In particular, sexual violence and threats of sexual violence are widespread but vastly underreported. The research team has acknowledged the prevalence of sexual violence in the accompanying commentary series.

Recoding actors

ACLED recorded 1,046 actors involved in political violence in Myanmar between 1 July 2020 and 31 May 2022. The research team classified these into four categories: ‘SAC forces’, ‘anti-SAC forces’, ‘ethnic armed organisations (EAOs)’, and ‘civilians’. Identification of EAOs using existing research on conflict in Myanmar was straightforward. Identification of ‘SAC’ and ‘anti-SAC’ forces was completed with the aid of local and international media reports. Non-armed actors engaged in violence were also classified into these categories if they had publicly declared their allegiance. Using this system, the research team reclassified all but 27 actors by the end of May 2022. The 28 events associated with these unclassified actors, comprising 0.21% of 13,431 events in the dataset, have been excluded from the map.

Unattributed violence

Between 1 July 2020 and 31 May 2022, the ACLED dataset recorded approximately 3,000 events in which an ‘unidentified armed group’ was involved. This corresponded to approximately 21.9% of all events in the dataset as of the end of May 2022. The unidentified armed group in question was classified as ‘SAC’ or ‘anti-SAC’ if they had attacked anti-SAC or SAC targets respectively. However, this was not possible in most cases. For instance, ACLED data records landmine explosions, bomb blasts in teashops, and abductions unclaimed by any actor. The ‘unattributed violence’ tab makes visible the spatial distribution of these known unknowns.

Loading map ...

Map of unattributed violence, “known unknowns”

Timeline

Jul ’20
Feb 2021
Military
coup
Jan ’22
Sep ’22
Violent
event

Updates to the map and release of additional analyses

The Myanmar Conflict Map will be updated by a team of researchers at the IISS on a regular basis as the IISS examines new data added to the ACLED database. Accompanying the map, analyses of six warscapes will be released periodically between now and September 2022. Additional analyses highlighting humanitarian issues and the relationship of Myanmar’s neighbours to the conflict will be published in the last quarter of 2022 and beyond.

Conflict cartography and its limits

The Myanmar Conflict Map strives for the best possible cartographic representation of armed violence in the country since the coup. In adapting ACLED’s data to the Myanmar context, the IISS sought to ensure that the vocabulary in the map speaks to ongoing policy debates about Myanmar, and that the map reflects the asymmetrical conflict between the SAC, anti-SAC forces, and EAOs. The research team also recognises the dangers that journalists and their sources currently face, and the SAC’s use of communications blackouts as a counter-insurgency strategy. These contribute to a significant incidence of unattributed and possibly unreported violent events. By rendering these unattributed events visible, the map recognises that the conflict in Myanmar is a struggle over information as much as a struggle over territory, resources and rights. The accompanying commentary series complements the map by addressing unreported violence and other forms of violence that do not lend themselves to cartographic representation.