Digital twin travellers: Disaggregated travel demand from aggregated mobile phone data – A privacy by design approach

Abstract

Mobile phone data generated in the mobile communication networks has the potential to improve current travel demand models and, in general, how we plan for better urban transportation systems. However, due to its high dimensionality, even if anonymised, there still exists the possibility to re-identify the users behind the mobile phone traces. This risk makes its usage outside the telecommunication network incompatible with recent data privacy regulations, hampering its adoption in transportation-related applications. Integrating mobile phone data in transport planning requires looking holistically at the data privacy implications within transport modelling systems.

In this thesis, the first Privacy-by-Design framework that integrates mobile phone data into disaggregated travel demand models and simulations, the Digital Twin Travellers, is introduced. Different strategies to modify generative Markov models for urban mobility plus an adaptation of the Rejection Sampling algorithm are explored to synthesise realistic individual travel demand from aggregated mobile phone data. In addition, the One-day Mobility Population Accuracy Score (OMPAS) and a Semantic Similarity (SS) score are introduced to measure the similarity between the generated digital twin population and the mobile phone user population, where mobile phone users are represented with their one-day itineraries.

The results show that by only requesting seven different types of histograms from a Telecommunications Service Provider (TSP), it is possible and plausible to disaggregate aggregated mobile phone data through the Digital Twin Travellers into new synthetic and useful individual travel demand data. Furthermore, the Privacy-by-Design approach ensures that mobile phone data can be easily shared in accordance with data privacy regulations between TSPs and transport modellers. Providing access to mobile phone data while protecting users’ privacy allows for streamlining the deployment of disaggregated transport applications such as multi-agent simulations for transportation and strategic planning in practice and ultimately unleashing the potential of big data for transport planning.