Mohammed Abdalazeem

Civil and Environmental Engineering, UMass Amherst

I’m a PhD candidate in Civil Engineering (Transportation Systems) at the University of Massachusetts Amherst, advised by Dr. Jimi Oke. My dissertation develops machine learning and optimization methods for designing schematic bicycle maps and understanding how map design shapes cycling behavior. I defended my prospectus in May 2026 and expect to graduate in fall 2026. I also hold an MS in Civil Engineering (2024) and an MS in Computer Science (2025), both from UMass.

My research combines transportation engineering, machine learning, and spatial analysis. On the transit side, I’ve built models that recover full passenger origin-destination patterns from boarding-only mobile ticketing data, the kind of incomplete data most transit agencies are working with. On the safety side, I use clustering and interpretable ML to understand how crash risk varies across neighborhood types in New England. Most recently I’ve been asking whether simpler bicycle maps actually work better at getting people to ride. That work spans survey design, optimization modeling, and behavioral analysis.

I have 8 published peer-reviewed articles, with 3 under review and 4 in preparation, in journals including Data Science for Transportation and Environmental Research Communications, and I present regularly at TRB and INFORMS. I’ve also mentored five undergraduate researchers, two of whom have published papers from their work with me. Before graduate school, I completed my BSc in Civil Engineering with First Class Honors at the University of Khartoum, Sudan.

news

Jul 21, 2026	Bot detection paper published in DST
Mar 01, 2026	Our collaborative paper “Scenario discovery framework aids robust regional emissions mitigation planning” is now published in Environmental Research Communications.
Feb 01, 2026	Schematic bicycle maps paper out in ERC
Oct 15, 2025	Presented a poster on “Optimizing and Deploying Schematic Bicycle Maps with MILP and User Input” at the INFORMS Annual Meeting in Atlanta.
Feb 28, 2025	Presented a poster on “A Roadway Crash Typology of Census Tracts Enables Targeted Interventions” at the NEUTC Annual Symposium at Norwich University, VT.

see all →

selected publications

ERC
Maps are good but are simpler maps better? Insights on urban bicycling in the US

M. Abdalazeem, G. Altayb, E. Christofa, and J. Oke

Environmental Research Communications, 2026

Abs DOI Bib

This research investigates cyclist preferences for map representations and route selection criteria through a national survey study. We analyze how schematic bicycle maps can influence mode choice and reduce vehicle emissions in urban areas, identifying key design features that affect cycling behavior.
@article{abdalazeem2026maps, title = {Maps are good but are simpler maps better? Insights on urban bicycling in the US}, author = {Abdalazeem, M. and Altayb, G. and Christofa, E. and Oke, J.}, journal = {Environmental Research Communications}, volume = {8}, number = {2}, pages = {025004}, year = {2026}, publisher = {IOP Publishing}, doi = {10.1088/2515-7620/ae3945}, }

DST
A Roadway Crash Typology of Census Tracts Enables Targeted Interventions via Interpretable Machine Learning

M. Abdalazeem and J. Oke

Data Science for Transportation, 2025

Abs DOI Bib

This study applies UMAP and Gaussian mixture models to classify 2,480 census tracts into crash typologies using state-level crash records in New England. Identified distinct spatial patterns in crash frequency and severity linked to roadway characteristics and socioeconomic factors. Built predictive models to forecast crash risk by typology, enabling targeted safety interventions.
@article{abdalazeem2025crash, title = {A Roadway Crash Typology of Census Tracts Enables Targeted Interventions via Interpretable Machine Learning}, author = {Abdalazeem, M. and Oke, J.}, journal = {Data Science for Transportation}, volume = {7}, pages = {14}, year = {2025}, publisher = {Springer}, doi = {10.1007/s42421-025-00128-2}, }

IJTST
Origin-destination inference in public transportation systems: A comprehensive review

M. Abdalazeem and J. Oke

International Journal of Transportation Science and Technology, 2023

Abs DOI Bib

This paper presents a comprehensive review of state-of-the-art methods for origin-destination estimation in transit networks, identifying key advancements, data sources, and modeling techniques. The review synthesizes research trends in OD modeling, emphasizing the transition from survey-based methods to big data-driven techniques utilizing automated fare collection, automatic vehicle location, and mobile ticketing data.
@article{abdalazeem2023origin, title = {Origin-destination inference in public transportation systems: A comprehensive review}, author = {Abdalazeem, M. and Oke, J.}, journal = {International Journal of Transportation Science and Technology}, volume = {12}, number = {1}, pages = {315--328}, year = {2023}, publisher = {Elsevier}, doi = {10.1016/j.ijtst.2022.03.002}, }

DST
Enhanced Seasonal Typology-Informed Transit Trip Chaining via Mobile Boarding and Survey Data

M. Abdalazeem and J. Oke

Data Science for Transportation, 2024

Abs DOI Bib

This study develops a trip chaining model incorporating passenger typology and seasonal variations to infer full passenger journeys using boarding-only AFC data. The model integrates a gradient boosting machine (GBM) for spatial error correction, reducing mean absolute error (MAE) by 70% and symmetric mean absolute percentage error (SMAPE) by 85%.
@article{abdalazeem2024enhanced, title = {Enhanced Seasonal Typology-Informed Transit Trip Chaining via Mobile Boarding and Survey Data}, author = {Abdalazeem, M. and Oke, J.}, journal = {Data Science for Transportation}, volume = {6}, pages = {19}, year = {2024}, publisher = {Springer}, doi = {10.1007/s42421-024-00108-y}, }

DST
Extracting Spatiotemporal Bus Passenger Trip Typologies from Noisy Mobile Ticketing Boarding Data

M. Abdalazeem and J. Oke

Data Science for Transportation, 2023

Abs DOI Bib

This paper develops a novel clustering framework using hierarchical clustering and dynamic time warping (DTW) to identify distinct spatiotemporal passenger typologies in the Pioneer Valley Transit Authority (PVTA) bus system. The method analyzes activation time series data from mobile ticketing transactions to infer behavioral patterns of different passenger groups.
@article{abdalazeem2023extracting, title = {Extracting Spatiotemporal Bus Passenger Trip Typologies from Noisy Mobile Ticketing Boarding Data}, author = {Abdalazeem, M. and Oke, J.}, journal = {Data Science for Transportation}, volume = {5}, pages = {20}, year = {2023}, publisher = {Springer}, doi = {10.1007/s42421-023-00082-x}, }