Using Open Data to Analyze Urban Mobility from Social Networks

Abstract

The need to use online technologies that favor the understanting of city dynamics has grown, mainly
due to the ease in obtaining the necessary data, which, in most cases, are gathered with no cost from social networks
services. With such facility, the acquisition of georeferenced data has become easier, favoring the interest and feasibility
in studying human mobility patterns, bringing new challenges for knowledge discovery in GIScience. This favorable
scenario also encourages governments to make their data available for public access, increasing the possibilities for data
scientist to analyze such data. This article presents an approach to extracting mobility metrics from Twitter messages
and to analyzing their correlation with social, economic and demographic open data. The proposed model was evaluated
using a dataset of georeferenced Twitter messages and a set of social indicators, both related to Greater London. The
results revealed that social indicators related to employment conditions present higher correlation with the mobility
metrics than any other social indicators investigated, suggesting that these social variables may be more relevant for
studying mobility behaviors.