Building a Visual Analytics Workflow for Location Based Services
Analysis of spatio-temporal data has become critical with the emergence of ubiquitous location sensor technologies and applications keeping track of such data. One potential application area is location based services (LBS) for mobile networks. In this research, we propose an interactive workflow to detect patterns and anomalies for quality control and customer behavior analysis. To form as a basis to our workflow design, we interviewed domain experts from a LBS company and identified different cases of LBS data analysis. Furthermore, we developed two visualization techniques namely density map and density cube techniques and conducted a laboratory experiment to compare aforementioned techniques from time and correctness perspectives. The results of our experiment revealed that density map technique leads to faster analysis and more accurate results for tasks related to anomaly detection in narrower time intervals and limited geographic areas whereas density cube technique is more efficient in terms of completion time and accuracy for tasks requiring a general trend analysis or spatio-temporal exploration of a given dataset window.
Dr. Selim Balcisoy received his PhD on Computer Science in 2001 from Swiss Federal Institute of Technology, Lausanne (EPFL). Between 2001 and 2004 he was Senior Research Engineer at Nokia Research Center Dallas. Dr. Balcisoy co-authored over 50 publications and has been granted with one U.S. patent. His research interests are augmented reality, data visualization and cultural heritage. He is an Associated Professor at Sabanci University, Istanbul since 2004 and founded VisioThink, a spatial business intelligence company in 2006.