Pipeline builds citation-grounded causal graphs from humanitarian reports

Researchers describe a processing pipeline that turns large volumes of heterogeneous humanitarian reporting into structured disaster storylines and causal knowledge graphs, with each extracted claim tied back to its source citation. The work targets the first hours of a crisis, when the amount of incoming information typically outstrips what human analysts can review. The approach is presented as a way to make automated synthesis of relief-related reporting more traceable and verifiable.