ManyCities maps urban residents’ daily lives to produce smarter insights

By September 21, 2015

A new freely accessible web application tool that analyses data from mobile phones reveals the distinct daily behaviour of residents of different cities.

The huge quantities of mobile phone data being collected nowadays can help to analyse and point up aspects of human activity which would otherwise be difficult to track. The many examples of such uses range from mapping daily commuter journeys to arresting criminals, helping to prevent famines, studying wealth distribution on the African continent , and even analysing couples’ personal relationships. Every such analysis calls for precise data to be carefully processed by experts. The data can however also be presented in a way that makes it accessible and understandable for a much wider public, rather like meteorological information. One example is the ManyCities project being carried out by Daniel Kondor and a team of researchers from the SENSEable City Lab at MIT and Swedish multinational communication technology and services provider Ericsson. The initial findings of the project, which will be shared and made accessible to everyone, have recently been published. The team has gathered and processed massive volumes of cellphone data from four major cities – London, New York, Los Angeles and Hong Kong – presented in simple and intuitive graphic form which anyone and everyone should be able to grasp. Among the major target audiences are local authorities, who should be able to use the ManyCities information to help them adapt existing urban infrastructure to the perceived needs of the local residents and/or with their future town planning.

Cellphone usage: wide variations according to place and time

The MIT-Ericsson researchers developed the ManyCities web application tool using several types of data drawn from the four metropolises over a period of ten months. This includes the number of calls placed, the number of text messages sent, the volume of data downloaded and uploaded and the number of data requests at 15-minute intervals. Interviewed by MIT Technology Review, the team confirmed that no personal details had been retained in the process: all data is anonymised. Data is presented in three different ways. Firstly, the study reveals how phone usage varies according to the time of day, highlighting daily and weekly patterns and also long-term trends. We learn for instance that city dwellers do not use their phones in the same way during holidays or major events such as the Wimbledon tennis tournament in London, and discover that data requests were rising constantly throughout the project period. Secondly, the findings underline how usage varies from city to city and within each city from district to district. For instance, the sending of SMS messages peaks during the morning in Hong Kong, at midday in London and in the evening in New York. Moreover, while people use their mobile phones more at the end of the day in New York and Hong Kong, the reverse is true for Londoners. In the Technology Review interview, Kondor explains: “We speculate that [this] is the effect of cellular data traffic being especially expensive in London, prompting people to switch to much cheaper WiFi networks when at home.” Thirdly, the data is looked at in terms of volume, revealing the most active parts of a city and how this varies over time and also allowing comparisons to be made across the world.

Facilitating production of developing country statistics

The ManyCities data is being made accessible to everyone. Local authorities in particular are likely to be able to use it to ensure that town infrastructure meets the current needs of the city or district residents. A perceived increase in mobile phone usage could for example indicate a need to expand public WiFi network coverage. In addition, organisers of private events such as conferences, sports contests and concerts could potentially use the data to select a venue where people are more ‘connected’ so that would have a greater opportunity to relay the event online. Early in 2014, a team of computer scientists in Paris were already using data from mobile phones to measure people flows across cities and concluded inter alia that in practically every urban environment in the world residents converge on the city centre during the day and depart again for the suburbs and outskirts in the evening. Other researchers stress the unrivalled potential of mobile data to help assemble statistics in emerging countries where local authorities often lack basic information on their own populations. This could be a simple population census or a means of assessing for instance the unemployment rate or the extent of an epidemic.

Legal mentions © L’Atelier BNP Paribas