Data and Information as Our New Transport Infrastructure: An Exploration into How the Modern Transport System Is Being Shaped by Information Communication Technology

This dissertation is focuses on the role that data and information has in creating and altering behavior related to transportation. To do so, it lays out a theoretical model of technological transition and then follows it up with three case studies. The theoretical model provides a structure to consider how different actors in our transportation ecosystem – users, firms, policy actors – mix with technological evolution to uphold or incrementally recreate our transportation landscape. The case studies stand on their own to highlight important findings about how data and information are impacting transportation scenarios, but collectively reinforce the theoretical models. The theory focuses on the idea of a socio-technical stack and the multi-level perspective of technological transition. The sociotechnical stack shows how a base of computing devices supply processing power that enables insight and then action. The multi-level perspective suggests that actors and influences sit at one of three interacting levels: niche, regime, or landscape. Niche actors are nimble and innovative but lack power; regime actors have power and influence but usually not speed; while the landscape is most often a set of conditions that require adherence or reaction to – sometimes imposed and sometimes created. Case 1: Big Data and Travel Desire: Comparing trip planner data exhaust to regional travel surveys The first case study shows niche actors in action. It focuses on the concept of understanding travel desire using firm or niche level journey planners. Travel desire is the need or determination to get from an origin or a destination before a trip has occurred. Data from journey planners where a user inputs their origin and destination to find a route is a source of travel desire data. This aggregated data is a byproduct of a service that users find valuable, as opposed to survey where the value is for the surveyor. This paper compares origin-destination data from a smartphone journey planner to the origin-destination data from Regional Household Travel Surveys (RHTS) for New York City and Philadelphia. The smartphone data has large, continuous sample sizes but lacks demographic information or sample controls to reflect the general population. The survey data has detailed demographic data and general population controls but is finite and thus best captures data for popular origin-destination pairs and modes on an average weekday. This study finds that at an aggregation level of two-or-more combined Zip Codes the smartphone and survey data show the same origin-destination patterns for trips where they both have data, despite their differences in collection and resolution. The smartphone data has the advantage of being continuous, widely dispersed, and has virtually no marginal collection cost over the core service of the app. Case 2: Taxis, Apps, and Transit: How the flow of information may redistribute transport supply to meet demand The second case study highlights the behavior and influence of regime actors in the face of changing technological conditions. Transport data that fuels smartphone-apps has progressively become a tool to help people achieve mobility by adding legibility, usability and reliability to the transport ecosystem. This cases goal is to understand the impact of data and information on transportation supply by evaluating the spatial distribution of New York City’s regulated and emerging for-hire-vehicle (FHV) market. By using increasingly robust data about vehicle assets, transport providers have found new ways to help match supply and demand. Here, data has two purposes: 1) to inform the traveling public of their supply options; and, if needed 2) to spatially match asset supply with user demand. This interaction has the possibility to shift supply or demand as users experience more options or operators seek under-served markets. Improved data reporting requirements coincided with new FHV-market entrants to form a natural experiment that reveals changes in transport supply. By comparing the spatial distribution of FHV’s in 2015 to a 2012 control, we see that supply increased in thinner markets in ways that are more complex than just adding supply to the street-hail system. This paper compares the spatial distribution of trip origins between the population of street-hail taxis, Uber, and Uber booked through a mobility-aggregator called Transit App to the 2012 control. It finds that as more segments of data & information are utilized to visualize or arrange supply, supply becomes more distributed relative to public transit service and the city core. Utilization of data & information appears key in helping supply to spatially distribute towards thinner demand. Case 3: Ridesourcing: friend or foe to transit? An exploratory study of overlapping supply in 5 US cities The final case highlights a landscape level issue: that technological transitions does not claim to happen evenly or without bias. As ridesourcing by transportation network companies (TNC) grow, there is interest in understanding how these rides are distributed across regions. A specific concern is that TNCs are siphoning users from transit. However, prior research also indicates that for-hire-vehicle (FHV) systems act as a compliment to transit by making first mile/last mile trips more reliable, allowing an alternative for high value trips, or providing a fail-safe for low-service levels, thereby extending the feasibility of transit. As ridesourcing expands the for-hire-vehicle market, it becomes consequential for transportation policy makers to understand how private ride-sourcing systems compete with mass transit. Using unique data from the Shared Use Mobility Center of ridesourcing origins in five cities, GTFS data from transit agencies, and the American Community Survey, this study contrasts the supply of TNC ridesourcing and transit systems at different times of day against area demographic profiles to understand their potential geographic consumer base. This exploratory analysis finds that while there is some demographic and spatial overlap, there is also a case for temporal complementarity. Transit is clearly designed to serve the most consistent market for mobility – high job density and high concentrations of car-less households at commute times – while TNCs also serve those segments, but with less predictability. In contrast, TNC origins are more likely to be supplied in areas with higher nighttime populations, a concentration of people age 25-44, and rising incomes. When the day is broken into different time periods, TNCs are found to have more variety in their demographic predictors than transit while also offering increased supply at times that transit supply is low. This dissertation then concludes that we are in a technological transition in mobility predicated the exchange and use of data. Through the mobility stack, algorithms routinely connect transport supply and demand, thereby enabling mobility. As highlight through mechanisms of the multi-level perspective, new services are regularly upsetting established systems. For example, one consequence of improved information in mobility is allow demand to dynamically attract supply. However, new systems also have biases. This transition in transportation should be understood and managed to get the best results