A comparative analysis of mobile phone data and travel surveys in understanding travel behaviour

Abstract This study evaluates the comparability of aggregated mobile phone data (MPD) derived from passive network signalling events and traditional travel survey data for urban transport planning, using the province of Liège as a case study. Our analysis demonstrates that while MPD captures a higher density of origin–destination (OD) connections, it cannot fully replicate all flows observed in surveys, underscoring the need for a complementary approach between the two data sources. Key mobility indicators, including average trip rates, hourly trip volumes, and structural patterns in daily OD matrices, show strong alignment. This structural similarity is rigorously quantified using a Mean Structural Similarity Index with a distance decay effect. Furthermore, Kolmogorov-Smirnov tests confirm comparable trip length distributions between the sources. While MPD-based population estimates closely match official 3:00 AM census counts, discrepancies in specific zones highlight potential pitfalls for real-time population mapping. Our findings confirm that MPD provides a robust and valuable complement to traditional surveys, particularly in contexts with limited or infrequent survey data. The study offers critical insights for integrating MPD into urban policy planning, emphasizing its utility for validation and its caveats for population estimation.