Unveiling Activity-Travel Patterns through Topological Data Analysis

TL;DR

This paper introduces a novel application of Topological Data Analysis to urban mobility data, revealing complex activity patterns and heterogeneity among residents, which can improve transportation planning and predictive models.

Contribution

It applies persistent homology to high-dimensional urban mobility data, enabling the extraction of complex activity patterns without data reduction, a novel approach in this context.

Findings

Identified five main activity-travel patterns among residents

Revealed behavioral heterogeneity linked to socio-economic attributes

Provided insights for urban planning and mobility prediction

Abstract

In the context of rapid urbanization, understanding the patterns of urban residents' activities and mobility is crucial for optimizing transportation systems and enhancing urban management efficiency. This study addresses the limitations of traditional travel analysis methods in handling high-dimensional and large-scale spatiotemporal data by incorporating Topological Data Analysis (TDA) techniques, specifically using persistent homology. This method allows for the extraction of information from the topological structure of data, enabling the effective identification and analysis of complex spatiotemporal behavior patterns without reducing the data's dimensionality. We utilized mobile signaling data from a community in Shenzhen, which includes detailed geographic and temporal information, providing an ideal sample for analyzing urban residents' behavior patterns. Using our pattern mining framework, we successfully identified five main patterns of residents' activities and travel, revealing daily behavioral habits and reflecting the activity heterogeneity among residents with different socio-economic attributes. These findings not only assist urban planners in better session design but also provide new characteristics for predictive mobility models.