Horizon CDT Research Highlights

Research Highlights

Understanding urban behaviour from mobile phone event series

  Gregor Engelmann (2014 cohort)   www.nottingham.ac.uk/~psxge

Abstract

Estimations show, that the majority of the world's population will live in urban areas by 2017. This will place a further demand on often already limited transport infrastructure. 35% of the 100 largest cities do not have complete transit route maps, with the number increasing to 92% in the 100 largest lower-middle income cities (Krambeck, 2015).

Traditional methods of data collection for the identification of travel and mobility trends ranging from GPS to other physical sensor technologies are not widely available in Lesser Economically Developed countries (LEDCs). The proliferation of mobile devices in recent years, however, has helped eschew a new era of human behavior monitoring at scale as devices as vast amounts of data are created at low costs. This Ph.D. will examine the potential role that Call Detail Records (CDR) can play in closing the data gap in data-poor areas around the globe.

The undertaken work involves the identification of what kind of data is used and at what levels of granularity to understand the requirements that CDR data needs to fulfill in order to be a viable alternative to traditional data collection methods. It will also use CDR data to understand different aspects of transport trends to showcase its potential viability in helping us understand mobility trends in LEDCs.

Potential outcomes will include methodological contributions of working with CRD, an improved understanding of mobility patterns in the case study area, and insights into the process of opening up confidential mobility data.

Background

In the past, numerous research projects and studies and been carried out using data obtained from cellular networks for various aspects of urban and transport planning. The investigated issues range from network data collection techniques (Valerio, 2009); qualitative and quantitative representation of the cellular networks data (Calabrese et al., 2011; Ratti et al., 2005; Ratti et al., 2006; Reades et al., 2007); travel time and speed estimation (Alger et al., 2005; Bar-Gera, 2007; Caceres et al., 2007; Herrera et al., 2010; Liu et al., 2008); correlation between cellphone traffic and vehicular traffic (Becker et al., 2011a; Caceres et al., 2007; Thiessenhusen et al., 2003; Vaccari et al., 2009); origin-destination estimation (Calabrese et al., 2011a; Iqbal et al., 2014; Pan et al., 2006; White and Wells, 2002); congestion detection (Hongsakham et al.,2008; Thajchayapong et al., 2006); incident detection (University of Maryland Transportation Studies Center, 1997); route classification (Becker et al., 2011a); inferring land use patterns (Becker et al., 2011b; Soto and Frías-Martínez, 2011; Toole et al., 2012); and inferring frequently visited locations (Ahas et al., 2010; Csáji et al., 2013; González et al., 2008; Isaacman et al., 2011).

Research questions

Since GPS and other physical sensor technologies are not widely available in Lesser Economically Developed countries (LEDCs), this raises the question, whether Call Detail Records (CDR) can be used to gain a representative insight into a city’s mobility patterns?

Are there fundamental differences and barriers in data application for, and in, transport planning use cases? Is it possible to detect urban activity patterns such as commuting hours from CDR data to better infer urban travel trends? What urban insights can be revealed from applying CDR data to dynamic topology mapping? What are the limitations of CDR data that affect its viability as a reliable data source for understanding travel trends and urban dynamics and how can they be overcome?

Research approach

In order to overcome these gaps in data provision, this Ph.D. will use CDR in order to gain an insight into general mobility and mass transit patterns. As working with CDR in mobility monitoring is relatively novel, this Ph.D. will use capability exploration using CDRs obtained from a major network operator in Africa.

  • Investigate the differences in data usage in transport planning in lesser economically developed and more economically developed countries

Information is required to effectively address issues and questions at strategic, tactical and operational levels. However, previous studies miss investigating the differences in data usage for different transport planning scenarios in MEDCs (More Economically Developed Countries) and LEDCs (Lesser Economically Developed Countries). An understanding of the differences is required to understand both gaps in data provision and application, and lessons learned in the integration of novel data types in an organizational usage context. One of the aims of this thesis is to investigate the differences in data usage (fidelity, requirements, collection methods, barriers), aggregation and visualization to identify data requirements for different transport planning tasks.

  • Investigate the potential of CDR data in detecting spatiotemporal distribution of urban travel

Traditional point detection, vehicle-based detection, and manual count data collection techniques are unsuited to most developing areas in the world due to a range of factors from fast land-use changes to prohibitive costs. CDR data, however, provides opportunities to overcome those issues. While previous studies have already addressed some planning scenarios in predominantly MEDCs using aggregated data, this research has the opportunity to model and infer using raw data over a lengthy time period. This includes the opportunity to detect local day-activity patterns instead of relying on western assumptions of activity windows. Thus, one of the aims of this thesis is to investigate effective ways for using CDR data to accurately predict localized urban travel demand patterns.

  • Investigate the role that CDR can play in understanding wider urban dynamics interrelated to transport and mobility trends

Ineffective data collection strategies and difficulties to retain an overview are not unique to transport systems alone. Nearly every aspect of urban areas from land-use to demographics and socio-economics is changing at too rapid a pace for conventional overview strategies. Thus, one of the aims is to investigate the feasibility of using CDR data for dynamic topology mapping.

  • Identify specifics and limitations of using CDR in mobility analysis and ways to overcome them

While previous studies have identified several issues in working with CDRs for mobility analysis, they have not had access to the same volume of raw data as is available as part of this project. During the first year of research, numerous issues that have not been discussed in the literature to date have already been identified. One of the aims of the Ph.D. is, therefore, a further investigation of limitations and specifics that need to be taken into account when analyzing CDR data.

Research contributions and outcomes

This PhD has the potential to make numerous methodological contributions in the areas of mobility mapping and urban information gathering. There are huge commercial opportunities for dynamic collection of mobility and transport information, particularly for fast changing environments such as rapidly growing cities and slums with high mobile phone penetration but a lack of funds/resourcing for formal GIS mapping.

The undertaken research can also contribute to an increased understanding of social and technical challenges associated with opening up confidential data sets. The underlying belief is that publishing data will encourage making it participatory and accessible, leading to innovation and benefit to the populace. The opening up of mobility data can lead to improved mobility information systems and encourage an increased adoption of public transport.

Supervision team

James Goulding - Horizon

David Golightly - Human Factors Research Group

Robin North - Transport Systems Catapult

References

  1. Krambeck, H. (2015). The General Transit Feed Specification (GTFS) and Implications for International Development. Presentation given at Transforming Transportation 2015, Washington, DC.
  2. Kenyon, S., Lyons, G. (2003). The value of integrated multimodal traveller information and its potential contribution to modal change. Transportation Research Part F: Traffic Psychology and Behaviour, 6(1),1–21
  3. Reitano, S. (2013). The Benefits of Open Data. Retrieved from http://www.beautifuldata.ca/Download/The_Benefits_of_Open_Data_-_Final_Report.pdf
  4. Bacon, J., Bejan, A. I., Beresford, A. R., Evans, D., Gibbens, R. J., Moody, K. (2008). Using Real-Time Road Traffic Data to Evaluate Congestion. In C.B. Jones, J.L. Lloyd (Eds.). Dependable and Historic Computing, (pp. 93-117). Berlin, Heidelberg: Springer
  5. Ehmke, J.F., Meisel, S., Mattfeld, D.C. (2010). Floating Car Data Based Analysis of Urban Travel Times for the Provision of Traffic Quality. In J. Barcelo, M. Kuwahara (Eds.). Traffic Data Collection and its Standardization. Volume 144 of International Series in operations Research & Management Science, (pp. 129–149). New York: Springer Science+Business Media
  6. Rose, G. (2006). Mobile phones as traffic probes: Practices, prospects and issues. Transport Reviews 26(3), 275–291
  7. Naranjo, J.E., Jiménez, F.J.S., Zato, J.G. (2010). Comparison between floating car data and infrastructure sensors for traffic speed estimation. Paper presented at IEEE ITSC2010 Workshop on Emergent Cooperative Technologies in Intelligent Transportation Systems, Madeira Island, Portugal
  8. Pu, W., Lin, J. (2008). Urban travel time estimation using real time bus tracking data. In Transport Chicago 2008
  9. Pu, W., Lin, J., Lon, L. (2009). Real-time estimation of urban street segment travel time using buses as speed probe. Transportation Research Record, 2129, 81–89
  10. Ding, N., Tan, G., Zhang, W., Ge, H. (2011). Distributed algorithm for traffic data collection and data quality analysis based on wireless sensor networks. International Journal of Distributed Sensor Networks, 2011.

Publications

  1. Engelmann, G (2016). Ethics and Privacy Implications of Using CDR Data for Social Good. Neodem Blog. Retrieved March 6, 2017 from http://neodem.wp.horizon.ac.uk/2016/10/12/ethics-and-privacy-implications-of-using-cdr-data/
  2. Engelmann, G (2016). Big Human Data: What are Call Detail Records?. N-Lab Neodemographics. Retrieved March 6, 2017 from http://neodem.wp.horizon.ac.uk/2016/10/06/what-are-call-detail-records/

This author is supported by the Horizon Centre for Doctoral Training at the University of Nottingham (RCUK Grant No. EP/L015463/1) and Digital Catapult, Transport Systems Catapult.