Leveraging principal component analysis to uncover urban pedestrian dynamics

Area of Science:

• Urban Planning
• Mobility Studies
• Data Science

Background:

• Urbanization increases city complexity, making pedestrian dynamics crucial.
• Pedestrian count data (footfall) offers insights into urban behavior but is complex.
• Understanding spatio-temporal structures of pedestrian flow is essential for city planning.

Purpose of the Study:

• To apply Principal Component Analysis (PCA) to extract temporal signatures from urban footfall data.
• To reveal dominant pedestrian activity patterns in urban environments.
• To analyze how PCA can inform understanding of urban functionality and mobility trends.

Main Methods:

• Utilized hourly pedestrian count data from Melbourne, Australia.
• Applied Principal Component Analysis (PCA) to reduce data dimensionality and identify key temporal patterns.
• Analyzed pedestrian volumes based on extracted PCA components.

Main Results:

• PCA successfully reduced dimensionality of noisy footfall data.
• Identified dominant patterns like weekday commuting and weekend leisure activities.
• Distinguished overlapping activities within locations and detected shifts in mobility trends, including impacts of external shocks like the COVID-19 pandemic.

Conclusions:

• PCA is a valuable tool for understanding complex urban pedestrian dynamics.
• The method reveals underlying urban activities and functionality.
• Findings highlight the utility of PCA for analyzing urban mobility and detecting changes over time.