Human Travel Exploration and Route Network Analysis using Campus Wi-Fi Log Data and Navigation API

Abstract: Understanding human travel activity and mobility patterns is a cornerstone for solving road congestion and optimizing roadway infrastructure design on university campuses. However, collecting data from a small-scale community, i.e. the university campus, for traffic modeling is challenging. For example, traditional travel survey approaches may be unsuitable for a university campus due to a lack of trip details and a limited sampling rate. Recently, widely deployed university Wi-Fi networks are logging their Wi-Fi users' communication activities, which provides an opportunity to collect human travel activity and understand human mobility patterns for campus-sized areas. For example, since 2021, the University of North Carolina at Charlotte (UNC Charlotte) has maintained and kept recording its campus-wide Wi-Fi log data, which contains details of Wi-Fi clients’ communication activities, such as connecting and disconnecting between mobile devices and access points (APs). Given the location of APs, communication activities imply human movement and travel patterns. In order to dig out human travel patterns from Wi-Fi data, this study proposed a novel “5-step” Wi-Fi log data process and analytics framework with navigation API that converts the raw Wi-Fi log data into the user travel activities or trips, including walking and driving routes. Basic statistical analysis and spatial and temporal analytic approaches have been applied to characterize user travel activities to discover human travel patterns and insights. For any origin and destination pair, the detouring behavior of both modes can be quantified by comparing the direct routes and the Wi-Fi routes through a Longest-Common-Sequence (LCS) method. Based on the detailed API-generated routes, road segments with heavy traffic can be ranked by a proposed “route-merging-and-split (RMAS)” approach, which doesn’t need detailed route GPS data and road network data. Wi-Fi-based human mobility patterns, detouring behavior, and ranked road segments can be used for various transportation applications, such as micro-transit or micromobility operations, campus travel demand modeling and surveys.