Mass movements and activities in public space increasingly pose major challenges to authorities and responders in terms of (1) availability and use of current and small-scale collective movement data for situational awareness and assessment, (2) analysis of data in near real-time, and (3) fusion of movement information (e.g. obtained from geo-social media, mobile radio data or in-situ image data). Such collective mass movements in public space have recently occurred and will occur more often in the future, e.g. in the course of behavioural restrictions in pandemic situations, demonstrations, large-scale events or refugee movements.

The MUSIG project focuses on the automated extraction of collective movement information from geo-social media, mobile phone data and in-situ imagery with AI methods, and the scenario-oriented fusion of movement information in a new mixed-methods approach, as well as its provision for crisis management and preven-tion in near real time including nowcasting information. In detail, the following methodological approaches are applied:

• Research into robust and transparent AI algorithms for multi-sensory analysis of movement information (number of persons, crowd distribution, density and behav-iour) from geo-social media, cell phone data and in-situ image data.
• Research into a mixed-method approach to fusing movement information: combina-tion of heterogeneous information layers.
• Legal, sociological and ethical issues are reflected in the technical research developments.
• The needs of end users are scientifically assessed in a structured manner and tested in a practical setting.
• Following a comprehensive exploitation strategy with a clear commitment to developing a joint service beyond the end of the project

By fusing movement information from heterogeneous sources, the creation of a high-quality and reliable basis of information, that can be used in a reliable way in crisis management and prevention, is aspired. Beyond the pure movement analysis (determination of number of people, density and speed of movement), the MUSIG project aims to extract semantic information (what do people in a group talk about?) and mood information (sentiment analysis – how relaxed, tense, escalating, etc. is the mood in a group). This collected information should serve as input for authorities, emergency services, security forces and rescue forces, etc. in order to be able to plan operations more precisely with regard to the disposition of personnel resources, the positioning of units, alarm response, the steering of flows of people, etc. The research results at will be tested on several geographical scales (urban, regional, national, international).