Hierarchical Framework for Processing Wi-Fi Log Data and Characterizing Community Human Mobility

Studies have shown that Wi-Fi data can be a reliable and qualified data source for human mobility analysis since it contains communication actions between clients and access points (APs) which can be used to infer human movement and travel activity. Understanding those human mobility patterns can be beneficial for developing effective urban management strategies. With more and more cities or communities providing public Wi-Fi service, a vast amount of Wi-Fi log data can be collected for human mobility analysis. However, Wi-Fi networks can have different network settings and log file formats since they were established using various internet equipment manufacturers and devices. Besides, bias and errors may exist in analysis results because of the complexity of Wi-Fi log data, such as ping-pong phenomenon and invalid messages. Thus, a common framework that can address public Wi-Fi data with different settings is needed to improve data processing efficiency and data analysis effectiveness. This study proposed a hierarchical and generalized Wi-Fi data processing and analysis framework to obtain and extract client activities from Wi-Fi log data. Wi-Fi log data collected from three areas: the University of North Carolina at Charlotte main campus, the city of Wilson and the town of Holly Springs in North Carolina, were processed and analyzed. Additionally, to obtain some insights into the correlations between Wi-Fi network settings and human mobility patterns, analysis results of human mobility patterns from the three study areas are examined and compared.