When travelling in the environment, people’s mobile phone periodically communicates with cellular network towers (even there are not call/SMS/internet activities), which leaves lots of “footprints” at these towers. These mobile phone network data allow us to potentially study travel behaviors of a high percentage of the whole population, with full temporal coverage at a comparatively low cost. However, extracting mobility information such as transport modes from these data is very challenging, due to their low spatial accuracy and infrequent/irregular temporal characteristics. For example, in the city like Vienna, the spatial accuracy of these data can be about 200-500 meters, while the temporal resolutions vary from 10 seconds to 1 hour.

This MSc project aims to develop methods to reconstruct complete trajectories from these sparse mobile phone network data. The proposed methods should be able to deal with mobile phone data with different spatial and temporal accuracies. Potential solutions might include map matching or (deep) machine learning.

Mobile phone data with different spatial and temporal accuracies will be provided. The corresponding GPS trajectories will be also provided as ground-truth to validate the proposed methods.

Methods, requirements: This project will focus on spatio-temporal data analysis. Skills in programming (e.g., Python or R), or willing to learn, are required.

Staff: Haosheng Huang (haosheng.huang@geo.uzh.ch) and Georg Gartner (georg.gartner@tuwien.ac.at).