Investigating Multiple Areas of Mobility Using Mobile Phone Data (SmartCare) in Chile

Abstract

Monitoring large scale mobility patterns is reliant on profiling the day-to-day movements of a significant number of city/country inhabitants. Mobile phone data (interactions with Telecommunication antennas) can be used to perform such profiling. In this paper, we present a program, Analysing Traces to Observe Mobility on SmartCare (ATOMS), to find and characterise user journeys. For Chile, we are able to profile more than 1 million users with approximately 3 million journeys/sub-journeys per day. For each journey/sub-journey, we find the start and end time, distance travelled, an estimate of the speed and further characteristics. Using the journeys stored in our database, Database of ATOMS (DATOMS), we are able to automatically identify commuters thanks to a second program, Neural Analysis of DATOMS for Itinerary Recognition (NADIR), by using a set of features from the journeys found by ATOMS in a Neural Network machine-learning approach. The potential for such a data-set is far reaching. We close by highlighting the potential (future) applications in mobility such as determining the mode of transport and inner-/intra-city Origin-Destination matrices.