Daytime Population Estimations Based on Mobile Phone Metadata

• Activity Number: 237
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Government Statistics Section
• Author(s): Martijn Tennekes*+ and May Offermans
• Companies: Statistics Netherlands and Statistics Netherlands
• Keywords: mobile phone metadata ; daytime population ; big data ; data science

Abstract

We provide a method to estimate the daytime population over time on municipal level with aggregated and anonymized mobile phone metadata. This method is developed in a pilot research collaboration project between Statistics Netherlands, Vodafone and Mezuro. To estimate the daytime population numbers, the Dutch population is split among several parts by age. For the part of the population that is assumed to use mobile phones, roughly between 12 and 75 years old, mobile phone metadata is weighted to the corresponding part of the Dutch population according to the Municipal Personal Records Database (MPRD). The numbers of children are estimated with primary school education registers. For the youngest and oldest part of the population, we only use the MPRD. We present preliminary results on the daytime population during one test month, May 2013, that contains spring holidays, Holy Thursday and Liberation Day. The results include daytime population peaks in the largest Dutch cities over time, daily commuting patterns, and classification of regions in working, commuting, and recreational. The cartographic plots and animations that are shown are made with the open source R-package geo.