Inferring modes of transportation using mobile phone data

Abstract

Cities are growing at a fast rate, and transportation networks need to adapt accordingly. To design, plan, and manage transportation networks, domain experts need data that reflect how people move from one place to another, at what times, for what purpose, and in what mode(s) of transportation. However, traditional data collection methods are not cost-effective or timely. For instance, travel surveys are very expensive, collected every ten years, a period of time that does not cope with quick city changes, and using a relatively small sample of people. In this paper, we propose an algorithmic pipeline to infer the distribution of mode of transportation usage in a city, using mobile phone network data. Our pipeline is based on a Topic-Supervised Non-Negative Matrix Factorization model, using a Weak-Labeling strategy on user trajectories with data obtained from open datasets, such as GTFS and OpenStreetMap. As a case study, we show results for the city of Santiago, Chile, which has a sophisticated intermodal public transportation system. Importantly, our pipeline delivers coherent results that are explainable, with interpretable parameters at each step. Finally, we discuss the potential applications and implications of such a system in transportation and urban planning.

Toward An Interdisciplinary Methodology to Solve New (Old) Transportation Problems

“…In this section we describe how the methodology was applied to the problem of inferring the mode of transportation share in Santiago (Chile) from anonymized mobile phone data [9,10]. This problem is traditionally solved through the collection of travel surveys, which provide rich information used to design changes and additions into the transportation network.…”

“…From this discussion we co-designed a visualization to compare OD matrices (see Figure 3). We applied this design to evaluate the OD matrices obtained for each mode of transportation (see reference [9] for its application to all modes of transportation).…”

“…The proposed methodology defines how both disciplines contribute to solve these problems through Algorithm Design, Data-Driven Evaluation, Interactive and Collaborative Evaluation, and Knowledge Consolidation. In this paper we describe each step of the methodology, including the actors and concepts involved, as well as a case study of a method to infer the transportation mode share of the population in Santiago (Chile) using mobile phone data [9].…”

Big Data Processing, Analysis and Applications in Mobile Cellular Networks

“…NASAs Tropical Rainfall Measurement Mission (TRMM) satellite ←→ anomalous patterns of mobility and calling frequency [38] Landsat-7 for deriving impact map of floods ←→ aggregated activity by day and by antenna [46] SPOT-Vegetation satellite for calculating vegetation index ←→ average number of calls between all market pairs [27] Environmental data The air quality estimated by regional model ←→ staying at home and travel patterns [17] Availability of environmental freshwater measured as the total length of the rivers in each spatial unit ←→ estimate of mobility obtained from CDRs [39] Logs of the climatic conditions: temperature, relative humidity, air pressure and wind speed from weather stations ←→ inferring the social network for each subject [49] POI Events on famous POIs across city ←→ users presences in the area [21] POIs from Google Earth for land use inference ←→ aggregated number of calls managed by each of base transceiver station towers [48] Pokémon POIs ←→ city-level aggregated distributions of number of connected devices and downloaded information from xDR records [24] IoT Inductive loop vehicle detectors ←→ mobility, rush hours traffic [28] Census, Surveys Travel surveys ←→ daily commuting from mobility traces patterns [3] Census on journey to work ←→ activity and connectivity around laborshed area [5] Demographic and health surveys ←→ connectivity and mobility across country [11] National statistics on socio-economic development ←→ human mobility patterns [45] Household income and expenditure survey ←→ top up credit amounts, mobility and social network features [57] Infrastructure The street network (highways and primary streets) from OpenStreetMap, metro network, bus routes ←→ xDR data aggregated into origin-destination (OD) matrices [23] Customer sites of each power line per grid square and line measurement indicating the amount of flowing energy ←→ aggregated people dynamics features from the mobile phone network activity [9] availability of external sources, efficiency of data processing and the quality of delivered information and its integration.…”

A Spatiotemporal Constraint Non-Negative Matrix Factorization Model to Discover Intra-Urban Mobility Patterns from Taxi Trips

“…Cazabet et al [34] proposed a method based on NMF to track the evolution of temporal patterns of usage in the bicycle-sharing systems in Lyon, France. Based on a topic-supervised NMF model, the distribution of the mode of transportation usage in Santiago, Chile were inferred from mobile phone data [35]. Maeda et al [36] detected urban changes through decomposing visitor arrivals by NMF.…”