How effective crowdsourced data during crisis emergency? A case of the 2018 Palu-Donggala earthquake

Abstract

In disaster situations, updated geographic data is crucial for disaster relief efforts. OpenStreetMap (OSM) has demonstrated significant value in disaster response scenarios due to its capacity for rapid data collection and dissemination, since the 2010 Haiti earthquake. This study investigates the quality of OSM data during the 2018 Palu-Donggala earthquake, focusing on how contributor expertise affects data reliability and how effectively OSM data supports decision-making in emergencies. The research highlights the critical role of OSM in providing timely geospatial information, with 205 contributors mapping roads and buildings in Palu City and Donggala Regency within just three days of the earthquake. Our findings show that while road data exhibited substantial topological errors—7,085 errors primarily due to overshoots—building data had considerably fewer errors, with only 76 recorded. This disparity suggests that OSM data for buildings was of higher quality during the crisis. The preference of eight out of nine mapper types for building data over road data further underscores the value of OSM in emergencies, as experienced mappers tended to focus on features that were less error-prone. The study also evaluates contributor behavior, revealing that while a significant portion of contributors were inactive, a majority of experienced contributors remained engaged. This finding indicates the potential for inactive expert mappers to return and contribute in future crises. Additionally, the study assesses the rapid collection of data by OSM and its impact on decision-making. The National Disaster Management Agency of Indonesia (BNPB) and the ASEAN Coordinating Centre for Humanitarian Assistance (AHA Centre) effectively utilized the data to provide updates on fatalities, injuries, and displacement, facilitating a swift and equitable distribution of aid.