Driven in part by the Covid-19 pandemic, the volume of online transactions globally has increased dramatically. While this presents business opportunities, it also represents a double-edged sword, as the cost to society of the growth in e-commerce activities has not been sufficiently evaluated for societal benefit. These costs - from the impacts on “brick and mortar” business outlets to changes in travel behavior, spatial mismatch challenges and the need to better manage public infrastructure such as curb access, are not insignificant. Very little data exists on the volume and scope of online deliveries, making it difficult if not impossible to evaluate the full costs, risks and social equity impacts of these changes. Exacerbating this is the changing nature of fulfilling these orders.

For example, micro-fulfillment is now the norm particularly for time-sensitive orders such as fresh groceries and cooked food, where these orders are often fulfilled at relatively high cost for last-mile deliveries. To address these issues, my research is developing a more holistic picture of e-commerce activities with publicly available microdata (US Census Public Use Microdata, Regional HTS, National and Regional Household Travel Surveys) and an approach that integrates both iterative proportional fitting (IPF), data imputation methods and machine learning.

The goal is to create a replicable and reproducible approach for estimating the demand of at-home delivery online purchases at granular geographical areas; by addressing last mile solutions.  In my research I am conducting natural experiments using Personal Delivery Devices (PDDs) with a focus on efficiency considerations on the task allocation/routing decisions and the energy implications and estimating the true cost to society of providing online delivery services. This work will include using a broad range of measures to assess the costs, including the shadow pricing of public assets (both free and subsidized) typically used in e-commerce activities.

This research strand examines the factors that determine individual travel behavior, with particular focus on the underlying elements that influence travel mode choices. Set against the backdrop of emerging mobility services, my work aims to deepen our understanding of how innovative mobility solutions can be systematically integrated into existing transportation infrastructure in ways that are both efficient and inclusive. A central challenge I address is the persistent issue of data interoperability within transportation systems that are highly complex, multidimensional, and comprised of numerous interconnected components. I also investigate the recurring disconnect between technological innovations and the diverse contextual realities where they are implemented, including the inherent tensions arising from conflicting objectives among multiple stakeholders.

Using socio-technical systems theory as my foundation, I am developing frameworks that go beyond conventional design principles to capture the dynamic interactions within complex urban mobility systems. A key component of this approach involves directly engaging affected communities in both the design and deployment phases of mobility innovations, thereby promoting inclusivity, legitimacy, and adaptive capacity. Alongside this theoretical and methodological work, my research explores how existing data can be repurposed or given new applications—for example, developing surrogate models to estimate vehicle cruising patterns without GPS data, or inferring on-street parking availability using smart meter transaction records.  

As a nod to the University of Michigan College of Engineering’s “People-first engineering” vision, I am conducting research on more robust operations research (OR) approaches for designing and implementing tech-enabled solutions that can advance the public interest in a more equitable manner. These activities are taking place on two fronts: I am integrating soft and hard OR methods to address the disconnect between technological innovations and the contextual realities of the environment in which these solutions are being implemented. I prioritize the voices of the community, with a particular focus on the populations of interest. These voices, elicited through community engagement, will be a key element of our research with the objective of surfacing and reflecting the impacted community’s viewpoint and preferences. My approach is to create a process where communities can influence and co-create, through a series of social interactions, tech-driven innovations in a manner that makes room for learning and that is transferable.

Secondly, the PIT space is inequitable, a consequence of an asymmetric power structure and lack of representation of historically marginalized voices. Many of the disputes with data and algorithm bias stem from these issues and from the simple reality that individuals tasked with developing solutions may have viewpoints or value systems that are not representative of the populations of interest. My research builds tools that will enable and encourage peer-to-peer mentoring among PIT practitioners of color that will incorporate their embodied knowledge and lived experiences and subsequently use the insights obtained to drive knowledge sharing, dissemination and learning within the community of PIT entrepreneurs. The goal is to create an environment in which tacit, localized and individual insights, knowledge and experiences could be converted to explicit knowledge, thus facilitating learning and knowledge transfer within the PIT community.