An Algorithm for Analyzing the City Residents' Activity Information through Mobile Big Data Mining

Abstract

The increasing of mobile devices results in the recent mobile big data era. A large number of useful information can be extracted from mobile big data. Extracting the residents' activity information from mobile big data is more and more popular in recent years because of its lower cost and higher accuracy. In this paper, we propose an algorithm to mine some meaningful residents' activity information from massive mobile data using Apache Spark. We first develop a weighted agglomerative hierarchical clustering algorithm to dig out the hot areas of a city and then the pedestrian flow of one hot area is analyzed. Next, we screen out those people who work in one hot area and dig out their residences and the destinations where they go after work and then the distribution situation of their residences and destinations are get based on the weighted agglomerative hierarchical clustering algorithm mentioned above. The results of this research reflect the city residents' activity information more authentic because the data we use is massive and is generated by the real activities of city residents.