When disasters strike, crowdsourced and citizen-reported location data, such as residents flagging flooded roads, blocked routes, or damaged areas, offers a fast, on-the-ground view of what is happening. But it arrives quickly and in large, noisy, often duplicated volumes.At the moment, responders and decision-makers most need a clear picture of where impact is greatest, that data is hard to trust and harder to prioritize. Converting dispersed, uneven reports into a reliable, severity-ranked view of the hardest-hit areas is a persistent challenge, and one that recent after-action reviews of major disaster events have repeatedly highlighted.
This talk explores an open, geospatial approach to validating and prioritizing crowdsourced disaster reports. It looks at how spatial clustering can consolidate duplicate reports into corroborated data points, how classification techniques can help filter unreliable inputs, and how density and concentration can be used to surface the most severely affected areas rather than simply the most frequently reported ones. The focus is on the underlying ideas and how they can be applied and adapted, rather than on any single implementation.
Along the way, the talk will share practical lessons from building and testing this kind of system — what works, where the approach reaches its limits, and how it has evolved over time. The code and data are open, and part of the goal is to invite others in the geospatial community to apply, extend, or improve on the approach in their own contexts.
The talk closes by looking at open questions in this space — particularly around validating these methods against real-world disaster data — and at how open-source tools and the FOSS4G community can help move the problem forward.
This session should be useful to anyone working with volunteered or citizen-reported geographic information, disaster and emergency-response data, or spatial data-quality and clustering methods. Attendees will leave with a clear picture of the approach, access to open code and data, and a sense of where the work is headed and how to get involved.