Comparative Analysis of Human Mobility Patterns: Utilizing Taxi and Mobile (SafeGraph) Data to Investigate Neighborhood-Scale Mobility in New York City

TL;DR

This study compares taxi and SafeGraph mobility data in NYC, revealing their respective strengths and limitations in capturing neighborhood-scale human movement patterns and transportation mode choices.

Contribution

It provides a comparative analysis of taxi and SafeGraph data, highlighting their differences in representing human mobility across NYC neighborhoods.

Findings

Taxi data captures high-demand areas like Lower Manhattan.

SafeGraph data better represents trips involving personal vehicle use.

Both datasets have biases in underrepresented neighborhoods.

Abstract

Numerous researchers have utilized GPS-enabled vehicle data and SafeGraph mobility data to analyze human movements. However, the comparison of their ability to capture human mobility remains unexplored. This study investigates differences in human mobility using taxi trip records and the SafeGraph dataset in New York City neighborhoods. The analysis includes neighborhood clustering to identify population characteristics and a comparative analysis of mobility patterns. Our findings show that taxi data tends to capture human mobility to and from locations such as Lower Manhattan, where taxi demand is consistently high, while often underestimating the volume of trips originating from areas with lower taxi demand, particularly in the suburbs of NYC. In contrast, SafeGraph data excels in capturing trips to and from areas where commuting by driving one's own car is common, but underestimates trips in pedestrian-heavy areas. The comparative analysis also sheds new light on transportation mode choices for trips across various neighborhoods. The results of this study underscore the importance of understanding the representativeness of human mobility big data and highlight the necessity for careful consideration when selecting the most suitable dataset for human mobility research.