Abstract
The percentage of processed large-scale heterogeneous data is exploding and technology is the most obvious reason for the big data issue. Nowadays, the results of data expansion are showing up in different fields. The users' contextual data are valuable in engineering and business domains, e.g., transportation, location-based services, and advertisement industry. Take mobile phones as an example. There are billions of subscriptions worldwide and sensor devices are digitizing people interactions. The data volume generated by mobile phones and the need to make better, fact-based, and real-time decisions, are the challenges facing researchers. Recently, new technologies based on cloud computing have emerged to process and analyze a large volume of data. We have utilized such technologies for the analysis of call detail records with the collaboration with a telecommunications company. We present an exploratory spatial data analysis algorithm and its analysis results. To prioritize different areas, detecting hotspots in a fast and accurate way is our objective. The findings of this research work can be helpful for urban planning and development as well as telecommunication infrastructure upgrading. Note to Practitioners -This paper proposes novel spatial analysis to detect the spatial distribution of mobile phones. It concerns the correlation of various spatial objects regarding their spatial and nonspatial dimensions to detect dependency among them. We have constructed a weighted neighbor list of spatial objects and implemented different autocorrelation tests regarding the variable of interests to compute spatial dependencies. By applying a Kernel Density method, we have identified the distributions. The results of density estimation demonstrate that the proposed approach is effective and feasible. Such analysis can help organizations better implement monitoring and evaluation plans at all levels and make the necessary infrastructure improvement to satisfy user requirements.