This talk is on the creation of the recently refactored and open-sourced Remote Fire Detection and Analysis Tool (RFDAT ). Using NASA FIRMS and ESA Copernicus Sentinel data, the RFDAT collects instances where there is potential malicious burning of civilian areas. It follows the methodology found here: https://www.info-res.org/sudan-witness/articles/how-sudan-witness-verifies-and-maps-sudans-burnt-villages/ The data is mapped here: https://felt.com/map/Map-of-Fires-in-Sudan-Sudan-Witness-khryT-Hdth-lHryq-fy-lswdn-shhd-lswdn-qJapVNDKS9BWUh0tJqYFWHD Big Data Republic have collaborated with Centre for Information Resilience to develop the RFDAT process to be more efficient and accurate than ever, allowing for the documentation of war crimes where ground access is impossible. We will cover: - The human impact of malicious arson of civilian areas in Sudan and how public documentation of such instances contributes towards greater awareness and accountability. - How we combined NASA's FIRMS and ESA's Sentinel-2 data to monitor large-scale destruction. - The pipeline we built using Python, DuckDB, Google Earth Engine, and SQLMesh to process, analyse, and visualize time-series satellite imagery. - How we deal with challenges like cloud cover, spatial resolution, and false positives in conflict contexts. - Lessons learned from building a robust and open source analysis pipeline that can be used by journalists, NGOs, and human rights monitors. This session offers practical insight into working with remote sensing data and shows how data engineering can be leveraged for humanitarian impact, far beyond simple dashboards and KPIs.

This talk is on the creation of the recently refactored and open-sourced Remote Fire Detection and Analysis Tool (RFDAT ). Using NASA FIRMS and ESA Copernicus Sentinel data, the RFDAT collects instances where there is potential malicious burning of civilian areas. It follows the methodology found here: https://www.info-res.org/sudan-witness/articles/how-sudan-witness-verifies-and-maps-sudans-burnt-villages/ The data is mapped here: https://felt.com/map/Map-of-Fires-in-Sudan-Sudan-Witness-khryT-Hdth-lHryq-fy-lswdn-shhd-lswdn-qJapVNDKS9BWUh0tJqYFWHD Big Data Republic have collaborated with Centre for Information Resilience to develop the RFDAT process to be more efficient and accurate than ever, allowing for the documentation of war crimes where ground access is impossible. We will cover: - The human impact of malicious arson of civilian areas in Sudan and how public documentation of such instances contributes towards greater awareness and accountability. - How we combined NASA's FIRMS and ESA's Sentinel-2 data to monitor large-scale destruction. - The pipeline we built using Python, DuckDB, Google Earth Engine, and SQLMesh to process, analyse, and visualize time-series satellite imagery. - How we deal with challenges like cloud cover, spatial resolution, and false positives in conflict contexts. - Lessons learned from building a robust and open source analysis pipeline that can be used by journalists, NGOs, and human rights monitors. This session offers practical insight into working with remote sensing data and shows how data engineering can be leveraged for humanitarian impact, far beyond simple dashboards and KPIs.