Divyakant Tahlyan

Northwestern University Transportation Center

600 Foster Street

Evanston, Illinois 60208

I am a PhD Candidate at the Northwestern University's Transportation Systems Analysis and Planning Program in the Department of Civil and Environmental Engineering. I am currently working at the Transportation Center as a graduate research assistant under the supervision of Prof. Hani Mahmassani. I am interested and have strong skills in analyzing data, building machine learning and econometric models and using these to understand and solve transportation problems.

Research Interests

Statistical, econometric and machine learning methods for transportation data analysis: discrete choice and ordered response models, integrated choice and latent variable models, count data models, heterogeneity and hierarchical models, structural equation modeling, latent class and latent transition analysis, machine learning based classifiers (SVM, neural networks, decision trees) and ensemble learning.
Social networks and travel: Egocentric network data collection and synthetic social network generation, social capital measurement, dimensions and impact on activity participation.
Route choice modeling: choice set generation algorithms, truck route choice, use of large scale passive data sources
Crash data analysis: frequency and severity modeling and prediction, zero-inflated and limited dependent variable models, unobserved heterogeneity and hierarchical models, spatial transfer-ability and temporal instability.
ICT and travel: role of perceptions, attitudes, experiences, constraints and satisfaction on tele-activities adoption and future retention, virtual/physical activity interaction, tele-activity and well-being.

news

Sep 4, 2021	I was awarded the Northwestern University Transportation Center’s `Dissertation Year Fellowship`.
Sep 3, 2021	Paper on characterizing visitor behavior using sequence alignment clustering is published. [doi]
Aug 24, 2021	Paper on simulating large-scale events using a network of heterogeneous queues is published. [doi]
May 2, 2021	I was awarded the 2021 `T&DI Bob Camillone Memorial Scholarship` by ASCE Illinois Section.

selected publications

TM

Characterizing visitor engagement behavior at large-scale events: Activity sequence clustering and ranking using GPS tracking data

Abkarian, H., Tahlyan, D., Mahmassani, H., and Smilowitz, K.

Tourism Management 2022

Abs HTML

This study uses GPS data of 1461 participants at a planned special event organized in Oshkosh, Wisconsin named AirVenture to characterize their spatio-temporal activity participation behavior. The GPS data is used to derive activity sequences for participants and study the attractiveness of various activities at the event site. A validation procedure is proposed using aerial photos, from which crowd density is estimated and compared to heatmaps of GPS data. A machine learning clustering approach is used to group participants into market segments on the basis of their activity sequences. The results show a prevalence of 6 behavioral groups with statistical tests confirming significant differences related to movement and time use. Finally, a multinomial logit model is formulated, demonstrating that age, prior visitation, and attendance plan (daily vs. weekly) affect the typological behavior. The results reveal valuable insights that can help special event organizers with related marketing and planning strategies.
TRR

Simulating Large-Scale Events as a Network of Heterogeneous Queues: Framework and Application

Cummings, C., Abkarian, H., Zhou, Y., Tahlyan, D., Smilowitz, K., and Mahmassani, H.

Transportation Research Record 2021

Abs HTML

Large-scale planned special events (PSEs) can pose unique transportation and logistics challenges. Data collection and simulation are important tools to address these challenges, although they are often difficult because of event size and complexity. This paper discusses methods to address the challenge of multimodal simulation at large PSEs through the context of AirVenture, a large week-long airshow organized by the Experimental Aircraft Association in Oshkosh, Wisconsin. Sampling and data collection techniques are discussed for a variety of modal processes like private vehicles, pedestrians, and shuttles, and for different situations like vehicle arrivals and departures, pedestrian queues, and shuttle systems. A flexible simulation framework for integrating these three modes and numerous activities is developed as a network of heterogeneous queues and queue-dependent choices. The simulation tested a variety of proposed policy changes around the site, including rerouting shuttle lines, and adjusting the system of vehicle arrivals to the site. Results of this study demonstrate the effectiveness and flexibility of the data collection and simulation methodologies. The techniques developed in this work can be used to improve planning and transportation systems at many other forms of PSE.
MTBG

An empirical assessment of the impact of incorporating attitudinal variables on model transferability

Tahlyan, D., Balusu, S. K., Sheela, P. V., Maness, M., and Pinjari, A.

Mapping the Travel Behavior Genome 2020

Abs HTML PDF

This chapter focuses on assessing the benefits of incorporating attitudinal and perception variables for the spatial transferability of travel forecasting models. Specifically, this study compares the spatial transferability, in an empirical setting, for three model structures: 1) multinomial logit (MNL), 2) integrated choice and latent variable (ICLV) models and 3) MNL with factor scores. Transferability is assessed by applying the models estimated in one spatial context to another spatial context. The data utilized for assessing the spatial transferability in the three contexts comes from a survey conducted among 1148 respondents across the United States – primarily the Midwest and Southeast. In the survey, respondents were asked about their preferred intended use of autonomous vehicles, along with personal and household characteristics, current travel characteristics, and perceptions about benefits and concerns related to autonomous vehicles. The study found that the ICLV models estimated had similar transferability to fixed coefficient MNL models with no improvement in transferability observed. But it was found that additional information that could lead to improvement of transferability was found when factor scores were directly incorporated into a MNL model. The chapter concludes with a discussion of possible transferability implications for ICLV model use.