Analysing Geo-social Media using Geospatial Machine Learning to Support Humanitarian Action

This project aims to develop multi-modal (spatial, temporal, semantic) machine learning algorithms for social media analysis to support humanitarian action through information provision in near real time. Therefore, this project covers the workflow from data gathering and filtering to analysing and classifying social media posts. Additionally, user-tailored visualisation of the results (e.g. in the form of a decision-support dashboard) is crucial to comprehensively and reliably support decisions in humanitarian action.
Recent experiences in our research demonstrate a broad need for up-to-date information products from a wide variety of actors, regional and national governmental institutions, the European Commission, as well as relief organisations and action forces, showing a clear market perspective beyond the end of this research project.