Visual Analysis of Mass Mobility Dynamics (MobilityGraphs)

The MobilityGraphs workflow addresses visual analysis of mass mobility dynamics through combined spatial and temporal simplifications. Starting from time-varying flow data showing people's presence in places and movements between them, the approach employs graph-based clustering to aggregate spatially close places with strong flows into regions. This spatial simplification reduces visual clutter while preserving large-scale movement patterns. Temporal clustering then groups time steps with similar flow patterns, reducing the number of distinct situations requiring analysis. Results are presented through coordinated views: a calendar showing temporal cluster distribution, cluster thumbnails displaying simplified flow graphs, difference views for comparing situations, and geographic maps providing spatial context. Interactive parameter adjustment enables exploration at various abstraction levels. The approach revealed routine mobility patterns in Greater London (mono-centric structure with morning centre-directed flows and evening outward movements) and Abidjan (polycentric structure with three activity centres), showing correspondence between flow topology and transportation networks.