Geographically Modified PageRank Algorithms: Identifying the Spatial Concentration of Human Movement in a Geospatial...

Area of Science:

• Geographic Information Science
• Network Analysis
• Human Mobility Studies

Background:

• Network approaches simplify spatial features into nodes and links, focusing on connectivity.
• Traditional network metrics like PageRank identify important locations but often ignore geographic factors.
• Geographic considerations such as proximity and location attractiveness are crucial for understanding spatial patterns.

Purpose of the Study:

• To propose two novel geographically modified PageRank algorithms: Distance-Decay PageRank (DDPR) and Geographical PageRank (GPR).
• To incorporate geographic factors into PageRank to better identify spatial concentrations of human movement.
• To evaluate the effectiveness of these modified algorithms in geospatial network analysis.

Main Methods:

• Development of Distance-Decay PageRank (DDPR) algorithm.
• Development of Geographical PageRank (GPR) algorithm.
• Comparison of DDPR and GPR with traditional network metrics using intercity and within-city human movement data.

Main Results:

• The proposed DDPR and GPR algorithms more accurately capture spatial concentrations of human residence than traditional metrics.
• Both algorithms demonstrated superior performance in both intercity and within-city settings.
• Distance decay was found to be a more significant determinant of human movement concentration than location attractiveness.

Conclusions:

• Geographically modified PageRank algorithms effectively identify key locations of human movement.
• Distance decay is a primary driver of human mobility patterns in geospatial networks.
• Geographic proximity remains a fundamental factor influencing human movement.