Quantifying and visualizing jobs-housing balance with big data: A case study of Shanghai

Existing jobs-housing balance studies have relied heavily if not solely on small data. Via a case study of Shanghai, this study shows how cellular network data can be processed to derive useful information, job and housing locations of commuters in particular, for those studies. Based on cellular network data, this article quantifies and visualizes Shanghai's jobs-housing balance with a much larger sample (n = 63 million), finer spatial resolution and greater geographic coverage than ever before. It identifies and geocodes the local commuters by Base Transceiver Station (BTS), which has on average a service area of 0.16 km(2). After detecting jobs and housing by BTS, it aggregates them by subareas of particular interest (e.g., traffic analysis zones, inner city, suburbs and exurbs) to local planners and decision-makers. It also visualizes the traffic flows associated with the actual (T-act), theoretical minimum (T-min) and maximum (T-max) commutes. It shows that Shanghai's commuting pattern is far from the extremes (indicated by T-max and T-min traffic flows) and Shanghai's relative balance of jobs with respect to housing is decent (3.2 km) despite its huge population (24 million) and land area sizes (6800 km(2)). The cumulative distribution of the Tact and Train flows vary more significantly when the commuting distance is less than 6 km. In theory, there is high concentration of both jobs and housing within a 6-kilometer radius across different locales of the city. This potentially allows over 95% of all the local workers to find a job within 6 km of his/her residence or vice versa. In reality, a much lower percentage (71%) of workers can enjoy such a benefit. This can imply that there is qualitative mismatch between jobs and housing. (C) 2017 Elsevier Ltd. All rights reserved.