Mobile phone data reveal hidden structural differences in Lisbon commuting

Mobile phone records from the Lisbon Metropolitan Area have been turned into detailed maps of commuting behaviour, and the results show that the region’s 18 municipalities differ far more in how people move than conventional origin–destination studies can reveal. The study, published in Applied Network Science, combined process mining with network science to analyse Call Detail Records covering 2.7 million reconstructed trips, focusing on morning commuting.

The researchers built case-centric event logs from the phone data and discovered Directly-Follows Graphs for each municipality, representing the sequences of trips people actually take. These graphs were treated as weighted directed networks and compared using eight complementary metrics, including degree-distribution Wasserstein distance, flow correlation, Jensen–Shannon divergence and a composite complexity index. The comparison uncovered systematic structural heterogeneity that aggregated origin–destination analysis cannot detect.

Lisbon city stood out as a structural outlier, showing the highest modal diversity and the most complex trip-sequence topology. Most peripheral municipalities, by contrast, were dominated by private car use. The analysis also showed that geographic proximity does not guarantee temporal accessibility: median commutes between neighbouring municipalities such as Sintra and Cascais exceeded 25 minutes, which the authors interpret as indicating infrastructure gaps.

When municipalities were grouped by the structure of their mobility processes, they formed spatially coherent clusters: a northern suburban cluster, a southern Tagus cluster and an outlying Lisbon cluster. These groupings were consistent across both structural and temporal metrics, suggesting that shared commuting patterns follow recognisable geographic bands rather than varying randomly from one municipality to the next.

To check the reliability of the approach, the team compared its inferred commuting patterns with 2021 Census data. The agreement was strong, with a Pearson correlation of 0.94 and a Spearman correlation of 0.90, supporting the validity of using mobile phone traces to reconstruct commuting flows at municipal scale.

The methodological contribution lies in showing that process mining and network science are complementary. Process mining yields structured event representations and captures modal sequencing, while network science offers robust tools for comparing structures across cases. Applied together, they expose the order and composition of trips, not just their endpoints, which matters for diagnosing accessibility inequalities.

The authors note an important limitation: the Directly-Follows Graph structures represent only mode–destination transitions, not full multimodal trip chains. Addressing that gap is left to future work. The article is open access and, at the time of indexing, was shared as an accepted manuscript carrying a permanent DOI, subject to further edits before the final Version of Record.

For urban planners, the findings suggest that two neighbouring municipalities can be connected geographically yet remain poorly connected in time, and that car dependence in peripheral areas produces mobility process structures distinctly different from Lisbon’s diverse, complex commuting web.

Subject of Research: Process mining and network analysis of mobile phone records to compare commuting structures across Lisbon Metropolitan Area municipalities

Article Title: From mobility traces to process networks: evidence of structural heterogeneity in the Lisbon Metropolitan Area

Article References: Filonchik, K., Pinto, J. D., Pinheiro, F. L., & Bacao, F. (2026). From mobility traces to process networks: evidence of structural heterogeneity in the Lisbon Metropolitan Area. Applied Network Science. https://doi.org/10.1007/s41109-026-00822-2

Image Credits: AI Generated

DOI: 10.1007/s41109-026-00822-2

Keywords: urban mobility, process mining, network science, Call Detail Records, Lisbon Metropolitan Area, commuting, accessibility, Directly-Follows Graphs, origin-destination analysis, municipal clustering, mobility, traces