The COVID-19 pandemic sparked a significant shift in how we work and shop, resulting in wide adoption of remote work and e-commerce. These changes led us to rethink various aspects of our lives, including our living environments, work practices, consumption patterns, and time allocation patterns, among others. The pandemic thus provides a unique opportunity to examine research questions surrounding the adaptability of human behavior and the potential for lasting change in the aftermath of such a global crisis. The primary aim of this dissertation is to gauge the impact of pandemic-induced flexibility on activity patterns, spatial habits, and schedule habits, exploring whether the changes observed during the COVID-19 pandemic represent permanent shifts or only temporary adjustments. I do this through: - Collecting a comprehensive, longitudinal national dataset tracking the multifaceted impacts of the pandemic on human behavior, attitudes, and beliefs, using a mix of passive and active data collection methods. - Proposing a new metric that captures individual schedule regularity over time, while accounting for specific day-of-week characteristics. - Developing an analytical framework that recognizes the multifaceted nature of impacts of the COVID-19 pandemic and its associated relaxation of spatio-temporal activity constraints on travel behavior, and distinguishes the nature of such impacts across activity patterns, spatial habits, and schedule habits. - Evaluating the impact of telecommuting, as a key characteristic of the lifestyle changes ushered by the COVID-19 pandemic, on time-use and the diversity of locations visited.

Methodological Contributions Human mobility has been repeatedly shown to be regular and predictable, as a result of both internal (e.g. circadian rhythms, psychological traits, sociodemographic characteristics, etc.) and external constraints (e.g. commuting requirements, social responsibilities, etc.). Such regularity has been shown to have significant impacts, such as increased social contact rates and playing a significant role in disease spreading. The relaxation of spatio-temporal constraints around key activities during the COVID-19 pandemic indicates a potential reshaping of such regularity and predictability. For example, as employees enjoy more autonomy on their preferred work environment and their schedules, including when to work and on what days to commute, it is reasonable to hypothesize that they would exhibit more irregular schedules, flexibly adjusting their activities to meet their own needs; running errands during regular business hours when work demands are not intense, and following working routines that might be synchronous with colleagues from different time zones. While intrapersonal variability in travel behavior is extensively researched by transportation researchers, such research has often addressed variability in mode use, trip frequency, distance traveled, or activity time use, leaving a missing gap in understanding schedule variability. Further, such research does not account for day-of-week characteristics, despite such consideration being important since outside societal obligations and constraints are usually tied to specific days of the week. In Chapter 3, I build on the extensive intrapersonal travel variability literature by proposing a new metric that captures the intrapersonal schedule similarity across weeks.This metric measures schedule regularity by computing the cosine similarity between the time allocation vectors of each individual for specific days of the week (i.e. Monday, Tuesday, etc.) across several weeks. In doing so, I control for characteristics of specific days of week, such as outside social constraints common to same day of the week. We use this metric to evaluate how the COVID-19 pandemic and the associated spatio-temporal activity constraints affected schedule similarity.

Empirical Contributions

DatasetTransportation researchers have tried to understand how the COVID-19 pandemic impacted different aspects of travel behavior. However, most of this research is cross-sectional, does not capture non-transportation factors that can influence travel behavior, and uses either active (survey) data or passive data, limiting our collective ability to understand the dynamic and interrelated impacts surrounding the pandemic. In Chapter 2, I present the design and implementation of a study aiming at the collection of data tracking the state of people throughout the COVID-19 pandemic in the U.S. I, along other collaborators, collected a rich panel dataset combining both active survey data and passive data from U.S. residents between January 2020 and September 2022. The fusing of the longitudinal active and passive data helps overcome the limitations of active or passive data when used individually and limitations posed by cross-sectional dataset and allows important research questions to be answered; for example, to determine the factors underlying the heterogeneous behavioral responses to COVID-19 restrictions imposed by local governments. The passive dataset overcomes provides a continuous stream of human mobility, compared to only location traces associated with cell phone activity, or use of specific applications, financial transactions, or transit services. This dataset complements existing datasets by: 1) combining large scale detailed passively collected data with a smaller subset of actively collected survey data, 2) designing a survey that covers broader aspect of participants life and behavior including personality traits, political views, and vaccination intention and status, 3) deploying multiple survey throughout the COVID-19 pandemic to overcome the limitations of cross-sectional studies, 4) deploying the survey to participants across the US, and 5) making our collected data accessible to other researchers. We acknowledge, however, that by being broader than other studies, we might not be able to capture deeper information on any singular aspect of human life during the pandemic. This dataset was the foundation of the remaining research presented throughout this dissertation, and supported other studies by several researchers.

Impact of COVID-19 on Activity Patterns, Spatial Habits, and Schedule HabitsOne of the key COVID-19 impacts was the relaxation of spatio-temporal constraints of key activities, notably work and shopping, allowing for greater flexibility in work hours, work locations, and consumption mediums. Such shift can disrupt the historically documented regularity of human mobility, opening up the potential for a wide array of impacts. These include changes in frequency, range, and time, as well as the diversity of travel destinations and schedules. Yet, much of the existing research has been limited to examining narrow aspects of the pandemic's impact on activity pattern, often considering only short-term impacts. In Chapter 3, I propose a framework that evaluates the impacts of the COVID-19 pandemic and its associated relaxation of spatio-temporal constraints around key activities on multiple aspects of activity patterns, and whether such impacts are temporary or permanent. I relied on well-documented metrics from both the traditional travel behavior literature (i.e., trip frequency, dwell-time, trip-timing) and ``mobility science'' literature (i.e., radius of gyration, location entropy, exploration rate) to evaluate the COVID-19 impact on activity patterns and spatial habits. I also proposed a new metric to measure schedule habits by computing the cosine similarity between the time allocation vectors of each individual for specific days of the week (i.e. Monday, Tuesday, etc.) across several weeks. The analysis results reveal a mixed picture; while some metrics have reverted to their pre-pandemic baselines, others have not. Regarding the schedule habits, we observe a paradox, that while large sectors of the workforce have shifted towards flexible work arrangements, schedule habits have strengthened rather than weakened.

Impact of Telecommuting on Out-of-home Nonwork Time-use and Diversity of Visited LocationsRecognizing the persistence of telecommuting well beyond the COVID-19 pandemic, I quantify the impact of telecommuting on time-use at out-of-home nonwork locations and the diversity of locations visited in Chapter 4. This analysis is motivated by the well-documented research that demonstrates the strong links between well-being and both the amount of time spent in out-of-home nonwork locations, and the diversity of such locations. I use quasi-experimental designs and the fused passive and active dataset described in Chapter 2 to control for unobserved individual confounders, thereby overcoming limitations of majority of existing research relying on cross-sectional observational data. Further, I build on current studies by assessing the impact of telecommuting on weekly time-use, moving beyond the conventional focus on daily time-use. I find evidence that during pre-pandemic, individuals spend an average of 114 minutes at out-of-home nonwork locations on telecommuting days relative to commute days, dropping significantly to approximately 64 minutes in the early phases of the pandemic, and slowly recovering to approximately 119 minutes post pandemic, or pre-pandemic levels, with the largest share (~85\%) of this time spent at discretionary locations. Further, I do not find evidence that this increase in time use on telecommuting days at out-of-home nonwork locations is additive at the weekly level, with this time being shifted from other days of the week. In terms of the diversity of locations visited, I find evidence suggesting that an additional day of telecommuting results in an average reduction of 0.35 in the number of unique out-of-home nonwork locations visited, split unevenly between discretionary (0.23) and maintenance (0.13) locations. These results contribute to the large body of evidence on the documented impacts of telecommuting on travel behavior and help further bridge the gap between travel behavior and causal inference.

SummaryIn conclusion, this dissertation aims to understand the dynamic impacts of the COVID-19 pandemic on activity and travel patterns. Methodologically, I build on the set of existing methods aiming to understand intrapersonal travel behavior variability by proposing a new metric that captures individual schedule regularity over time. Empirically, we collect a unique national, longitudinal, dataset mixing both passive and active data collection methods tracking the state of people throughout the COVID-19 pandemic. I analyze this dataset to evaluate the impacts of the COVID-19 pandemic on activity patterns, spatial habits, and schedule habits, relying on well-documented metrics from the travel behavior and mobility science bodies of literature. Finally, I use causal inference methods to evaluate the impact of telecommuting on workers time use and diversity of locations visited.

Findings from this research reflect the dual nature of travel behavior following the COVID-19 pandemic, illustrating a balance between behavioral inertia and adaptability to a new post-pandemic world. First, within the framework proposed in Chapter 3, I find that while some mobility metrics have returned to their pre-pandemic baselines as early as 2021 (Weekly trips, Radius of Gyration, Share of peak trips, exploration rate) as people started emerging out of pandemic-induced lockdowns, other mobility metrics have yet to reverse to pre-pandemic baselines (At-home dwell time, location entropy), with people spending on average more time at home and visiting a less diverse set of destinations. Perhaps most interestingly, we find that despite the relaxation of spatio-temporal constraints of key activities, schedule habits have strengthened as opposed to weakened. This finding is striking, going against our initial hypothesis, and possibly that of many other researchers, that the flexibility induced by the COVID-19 pandemic will result in people exhibiting less regular schedules. Finally, in my analysis of telecommuting, I find that while telecommuting results in an increase in time-use at out-of-home nonwork locations on telecommute days, relative to commute days, such increase only represents a shift of time allocation at out-of-home nonwork locations across days of the week. Going beyond time use, I also find that telecommuting results in workers visiting less unique locations. Collectively, the findings reveal a complex picture of how the COVID-19 pandemic, and its associated relaxation of spatio-temporal activity constraints, has impacted activity patters, spatial habits, and schedule habits, with such impacts being neither fully temporary nor completely new.

The purpose of this review is to present findings from recent research on Shared automated vehicles (SAV) impacts on mobility and energy.

While the literature on potential SAV impacts on travel behavior and the environment is still developing, researchers have suggested that SAVs could reduce transportation costs and incur minimal increases in total trip time due to efficient routing to support pooling. Researchers also speculate that SAVs would result in a 55% reduction in energy use and ~ 90% reduction in greenhouse gas (GHG) emissions.

SAV impacts on mobility and energy are uncertain. Researchers should carefully track SAV technology developments and adjust previous model assumptions based on real-world data to produce better impact estimates. SAVs could prove to be a next technological advancement that reshapes the transportation system by providing a safer, efficient, and less costly travel alternative.