“A New Geography of Civil War: A Machine Learning Approach to Measuring the Zones of Armed Conflicts.” Political Science Research and Methods 10(1): 97–115. For details of the variables in the summary data, refer to the codebook of the UCDPGED. Finally, ged_summary.csv provides basic information of each conflict dyad, which can be merged to the conflict zone data by using dyad_id variable. The static.zip contains time-invariant estimates of conflict zones and their lower and upper confidence intervals. The shape files have three columns 'dyad_id' (new dyad IDs in the UCDPGED), 'conf_id' (new conflict IDs in the UCDPGED) and 'date' (date of the conflict zone). Each zip file contains shape files (365 or 366 files daily).

There is a separate zip file for each year (1989-2021). A Youtube video is also available at here. The dataset is available in the ESRI Shapefile format for each conflict dyad at a daily level. The conflict zones are created from the UCDPGED event data (version 22.1) with a machine learning method, so-called a one-class support vector machine, which is detailed here. Update (): The dataset is updated to UCDPGED 22.1.

The armed conflict database