The rise of technology-enabled ride-hailing services, like Uber and Lyft, have affected travel behavior and transportation-related decisions of individuals, which in turn has consequences for environmental sustainability. However, the impact of these ride-hailing services is likely to be heterogeneous, varying across traveler segments that differ in terms of their usage of various modes of transportation. Policy decisions based on analyses that do not consider why we observe impacts of ride-hailing services on transportation decisions, i.e. by considering the varying impact on the traveler segments, may be misguided. The goal of this paper is to examine the impact of ride-hailing services on sustainability, by examining the impact of Uber on the transportation mode choice for three different traveler segments: drivers (who primarily use personal automobile to travel), riders (who primarily take public transit to travel) and walkers (who primarily take non-motorized modes of transport). To do this, we first develop a framework that outlines how the behavior of these different traveler segments will be impacted by the presence of a ride-hailing service. Next, we compile a rich dataset, that includes data on public transportation ridership, the degree of traffic congestion, and survey data on individual transportation mode choice. Then, we employ difference-in-difference methodology to estimate the impact of Uber, the dominant ride-hailing service, on the role of each traveler segment in increasing/decreasing traffic congestion and public transit ridership. Our results show that Uber entry has allowed walkers as well riders to utilize Uber to travel more conveniently as reflected in an increase in traffic congestion, while inducing drivers to substitute their use of private automobile with a combination of Uber and public transit. By estimating our models using traffic congestion as the dependent variable and supplementing this with analysis using public transportation ridership data, we are able to get at the nuances of the changes in transportation decisions by each of these traveler segments. We also introduce urban compactness as a measure that can be used to assess the heterogeneous impact of ride-hailing services for different cities which differ in terms of the distribution of the traveler segments. Our findings confirm that the entry of Uber increases traffic congestion and reduces public transit demand more in a city with a higher urban compactness index, i.e., where the proportion of riders and walkers is higher than that of drivers. Our results, which allow us to understand the underlying mechanism driving observed impacts on traffic congestion, helps explain the inconsistent findings in the extant research. Further, urban planners can leverage our conceptual framework, methodology, and empirical results to guide city planning decisions based on the proportion of different traveler segments in a given city, and policy makers can use our results to examine implications for sustainability.