Large scale geospatial analysis on mobile application usage

Abstract

The recent exponential growth of mobile application instances in combination with the availability of more advanced networks have led to a significant increase of the usage of mobile devices and applications. Several studies regard application usage location and time as strong contextual characteristics and infer that user mobile usage habits can be affected by the user's location, such as rural areas and points of interest (schools, airports). In this work we consider a novel approach of collecting usage information from mobile devices, correlating it with other data sources such as Open Street Map and integrating it in spatiotemporal data warehouse. From there we can identify signatures of such usages and the impact of usage time and location. This experimental study includes an analysis of usage logs gathered from thousands of users spread over 127 different countries in a yearly span and presents several extracted correlations and trends such as we detect usage pattern and traffic trends during holiday periods such as increased usages in villages and lower ones in big cities.