Determining Spatial Co-Location Graph Influencing Factors using Mobile Applications Data

Abstract

With the recent explosion of smart phone users and the usage of mobile applications, it is undeniable that mobile applications are rich data sources for data mining. In the recent studies conducted by Statista, for 2016, the number of smart phone users is forecast to reach 2.1 billion while the number of mobile phone users in the world is expected to pass the five billion mark by 2019. The ownership of smart phones also leads to the rise of mobile application usage where most of the mobile app usages help users to manage life activities, increase productivity, and constantly stay connected while on-the-go. With permission granted, some mobile applications provide devices location with reasonable high accuracy. Spatial information derived from these mobile applications allow better understanding of human lifestyle behaviors as well as determine the social influence patterns. In this paper, we investigate factors that contribute to spatial colocation graph data mining patterns for the period of three months data (October to December 2018) for three SouthEast Asia countries, namely Malaysia, Singapore and Thailand. Devices that are co-located are defined by determining two or more devices are appeared in close proximity within a predetermined range of distance at the same time frame. Based on these co-located devices, spatial co-location graphs were constructed to analyze factors that influence graph.