How land use and access to transit impact taxi demand

By Rayla Bellis

Significant research and debate in recent years have surrounded the impacts of ride-hailing services like Uber and Lyft on transportation systems: whether they reduce the need for personal vehicles, how they contribute to or reduce congestion, and how they impact transit ridership. SSTI wrote recently about Chicago Mayor Rahm Emanuel’s proposal to increase the city’s fee charged to ride-hailing companies to offset the anticipated loss of revenue from public transit users who switched to ride-hailing services.

A recent study published in the Journal of Transport Geography may help shed further light on some of these questions by examining taxi demand and its correlation to land use patterns and access to other travel modes in the Washington D.C. region. As the researchers point out, despite the significant growth of on-demand ride-hailing service providers like Uber and Lyft, taxis remain a key asset for urban mobility that can either complement or compete with other modes. An improved understanding of the factors that impact taxi demand could not only lead to better policy decisions, but also help provide insights into how growing on-demand services like Uber and Lyft are likely to perform and impact other modes. This could help agencies and policy makers better integrate these emerging modes into multimodal transportation networks.

To analyze taxi demand, the researchers collected pick-up and drop-off times and locations from a one-year taxi ridership log from May 2015 to April 2016 provided by the District of Columbia Department of For-Hire Vehicles. The dataset includes all public vehicle for-hire trips (both dispatched and hailed) in DC and surrounding cities and counties through the reciprocity agreement with these jurisdictions.

The study team then integrated the pick-up and drop-off data with land use and transit network data using ArcGIS. In order to relate taxi demand with land use characteristics, which were only available at the Traffic Analysis Zone (TAZ) level, the study aggregated taxi demand for each TAZ based on the pick-up and drop-off locations. The researchers noted that unlike previous studies, which have relied primarily on measures of density to capture land use, this study also included measures of connectivity, centrality, level of mixed use, and accessibility to various destinations to more fully account for the breadth of urban form characteristics that might impact taxi demand.

To measure accessibility to transit, the researchers collected the geolocations of the Metro stations and bus stops in the entire metropolitan area from the DC GIS Open Data Site and the General Transit Feed Specification website. They also integrated data on time of day and seasonality to account for other factors likely to impact taxi demand.

Overall, the results of the study show a strong link between taxi demand, land use patterns, and accessibility to other modes in the DC area. Notably, while the findings show clear relationships between taxi demand and access to transit, that relationship changes depending on the type of transit service. The study found a positive correlation between taxi demand and the number of metro stations within a half-mile buffer of each TAZ, while the number of bus stops contributed negatively. This indicates that taxi transportation is likely to complement Metro service—potentially by providing first-and last-mile connections—but compete with bus service. The researchers suggest that these findings could be related to the income level of Metro riders versus bus riders, but note that additional research would be needed to test that hypothesis.

The findings also suggest that taxi demand is significantly higher in areas with an unbalanced mix of land uses, such as major commercial centers, central business districts, and residential neighborhoods. The authors note that from this perspective, taxi demand differs from the demand for transit, although both tend to be higher in downtown areas where activities are concentrated. They also note that this correlation only existed for the pick-up locations and not the drop-off locations, suggesting that the level of land use mix of the trip destination is less relevant to demand.

These insights could help transportation agencies and urban policy-makers make decisions about the transportation networks they manage. As the researchers note, comparing taxi demand with other on-demand services like Uber and Lyft will be another crucial piece of the puzzle for future analysis.

Rayla Bellis is a Program Manager at SSTI.