VARICRIS data set

Measuring and comparing transboundary crises

To facilitate theory-building and theory-testing in the project, as well as to enable future cummulative research by the community, VARICRIS will provide an original large-scale dataset that offers systematic information on the intensity, temporal dynamics, as well as the geographical and functional transboundariness of crises perceptions across the globe and across time (1994-2025).

To this end, we will build on an original text corpus covering the news stream of the largest leading international newswire agencies (which will also available for individual qualification projects of the team members), to then extract the relevant crisis features with a targeted NLP measurement model.

The exact methodological approaches can and will be sharpened together by the whole project team, but an initial proof-of-concept study presented by both PIs at the DVPW conference in Göttingen 2024 underlines the feasibility of the sketched approach. Two crisis case studies from the initial data set are illustrated below.