Urban mobility analysis of megacity New York using density-based clustering techniques

Abstract

This study presents a comparative analysis of two density-based clustering algorithms, such as DBSCAN (Density-Based Spatial Clustering of Applications with Noise) and HDBSCAN (Hierarchical DBSCAN), to evaluate their performance in identifying taxi demand patterns. Using the New York City Taxi and Limousine Commission’s (TLC) September 2015 yellow taxi dataset, which includes variables such as trip distance, duration, fare amount, passenger count, and geographic coordinates, the research applies clustering and correlation analyses to uncover spatial and behavioral mobility insights. The dataset was preprocessed to remove noise and inconsistencies before implementing both algorithms. Results show that HDBSCAN outperforms DBSCAN by effectively detecting clusters with varying densities, particularly in central Manhattan, and providing clearer, more accurate representations of taxi demand. In contrast, DBSCAN classified a larger proportion of the data as noise, revealing its limitations when the data are heterogeneous and highlighting its limitations in handling heterogeneous spatial distributions. Complementary analysis using Pearson’s Product-Moment Correlation Coefficient revealed a weak positive relationship (r = 0.21) between trip distance and duration, indicating the influence of traffic and route variability, and a strong correlation (r = 0.92) between trip distance and fare, reflecting the city’s distance-based pricing structure. The findings demonstrate that HDBSCAN offers superior performance for analyzing large-scale transportation datasets. The study provides empirical evidence supporting its use in urban mobility research and practical insights for transportation planners and policymakers to maximize fleet distribution, manage demand, and improve service delivery. Future work should integrate contextual variables such as time of day, payment methods, and tipping behavior to achieve a more comprehensive understanding of urban transportation dynamics.