Operational Data Science Lab
Lab Overview
Many functions of university infrastructure, like transportation and utilities, routinely collect data during their operation. Texas A&M University has over 21,000 employees and 77,000 students who are constantly using and creating data through research, education, and supporting our nineteen colleges and schools. This data is used to manage these functions from planning to daily operations across Texas A&M’s large and diverse campus.
The Operational Data Science Lab is a joint enterprise between TAMIDS and the Texas A&M Center for Geospatial Science, Applications, and Technology (GEOSAT) to develop partnerships with operational organizations in Texas A&M that capitalize on institutional data investments to improve campus operations, achieve research and academic goals, and support Texas A&M researchers working on operational problems.
- Identify new data-driven research problems arising from operational contexts
- Derive new insights and analyses from data and embody these in software tools
- Engage the broader community of researchers with operational data challenges
- Provide hands-on opportunities for students to work with real-world data
- Leverage the results for funding proposals and outreach
Achieving these goals has the potential to transform Texas A&M into a living laboratory for Data Science, accelerating the pace and effectiveness of Texas A&M’s teaching, research, and outreach mission. We would like to thank our partner operational organizations, including Texas A&M Transportation Services, Texas A&M Transportation Institute (TTI), University Libraries, Utilities and Energy Services, Division of Research, Texas A&M – Galveston, the Department of Animal Science, and the Department of Biology.
Members
Dr. Yalong Pi
Operational Data Science Lab Director Division of Research TAMIDS Associate Research Scientist Division of Research TAMIDS
Dr. Xinyue Ye
Urban Artificial Intelligence Lab Director Division of Research TAMIDS Professor College of Architecture Landscape Architecture and Urban Planning
Aniruddh Adiga
MS in Data Science Division of Research TAMIDS
Snehashis Chakraborty
Research Engineer Division of Research TAMIDS
Dr. Nick Duffield
Director Division of Research TAMIDS Professor Department of Electrical and Computer Engineering
Dr. Jian Tao
Digital Twin Lab Director Division of Research TAMIDS Assistant Professor College of Performance, Visualization, and Fine Arts Visual Computing and Creative MediaProjects
This work is pioneering the use of computer vision to transform data collection and analysis in livestock management. The project focuses on developing and refining algorithms to accurately identify and monitor animal behavior and health indicators. Using supervised, unsupervised, and statistical tools, the project has successfully demonstrated high performance in cattle behavior analysis through video monitoring, indicating a significant step forward in precision livestock farming.
Research Areas: Computer vision, Image classification, object tracking, motion detection
Traffic and Pedestrian Data Analysis via Video Monitoring: Central to this project is the utilization of advanced computer vision techniques to analyze vehicle (boxes) and pedestrian (points) flow. We employ Convolutional Neural Networks (CNNs) for object/point detection and tracking algorithms for assigning detection IDs and traces. These methods were instrumental in analyzing traffic patterns around Texas A&M football games, offering insights for real-world traffic analysis to TTI and TAMIDS Transportation Service.
Research Areas: AI-based object classification; motion tracking; transportation
This project integrates deep learning with social media data analysis for prompt damage estimation in disaster scenarios. By employing a deep neural network (DNN) that processes OpenFEMA data and geotagged social media content, we’ve developed a model capable of performing feature extraction through natural language processing and computer vision. This model was effectively tested during Winter Storm Uri, demonstrating the potential of machine learning in enhancing disaster response and resource allocation strategies.
Research Areas: Social media analysis, geospatial data analysis, image classification
This project aims to gather Land Mobile Radio (LMR) data from a wide range of public safety organizations (PSOs) across the country. The primary objective is to conduct research and analysis to gain insights into the potential impacts of transitioning from LMR to LTE technology. Our project team is collaborating with ten different PSOs nationwide, acquiring not only LMR data but also relevant geographical details, including weather and event information. These data sets will serve as resources for in-depth research and analysis, allowing us to better understand LMR usage patterns and to prepare for a successful transition to LTE.
Research Areas: Mobile Radio Technology, GIS, Public Safety
This work uses information concerning faculty publications, grant proposals, and other data to profile Texas A&M’s research in thematic research areas and to identify unexplored opportunities for future research. This project will develop and use software tools that enable exploration and characterization of current and trending research, visualize research collaborations through coauthorship networks, and identify latent and emerging opportunities for collaboration through subgraph prediction.
Research Areas: graph analysis, graph clustering, link prediction, keyword analysis
Tackling challenges in the rapid identification of C.difficile spores, which is essential for understanding the infection process, our lab is developing a point-based deep learning computer vision system. This system is geared towards automating the 3D localization (adhered and internalized) process of the spores in confocal microscopy, especially small objects (500-1000 nm). This tool will significantly speed up the research processes of devising targeted treatments, monitoring treatment effectiveness, and preventing recurrent infections.
Research Areas: Computer vision, image segmentation, machine learning, image classification
Universities have a responsibility to identify and regulate research in areas such as biosafety, animal subjects, and human subjects. This project developed a tool to assist research compliance officers by alerting to “red-flag” areas at the proposal stage. The tool used Natural Language Processing (NLP) to enhance proposal review by generating flag categories from proposal documents, and indicating text associated with specific red-flag categories.
Research Areas: Natural Language Processing, Text Processing, User-Interfaces
In collaboration with the Texas A&M Utilities & Energy Services (UES) and the Department of Statistics, this project is to analyze the time series data from the measurement of energy consumption of a building on campus and automate the process to detect anomalies. The framework created in this project is able to help the UES detect outliers and help them with their resource allocation decisions as well as maintenance schedules. This work is based on Prophet, an open source project by Facebook to forecast time series data based on an additive model where non-linear trends are fit with yearly, weekly, and daily seasonality, plus holiday effects.
Research Areas: time series analysis; time series forecasting; outlier detection
This work uses the real-time locations of TAMU rideshare bikes (starting with Ofo and now using VeoRide) and their paths to measure the effectiveness of the bikeshare program and assist with the operational rebalancing efforts to ensure that bikes are where users need them, when they need them. This project will develop software tools for data visualization and exploration, automated processes and database structures for data ingestion, storage, and retrieval, and geospatial machine learning models and approaches for predictive analytics to guide operational efforts.
Research Areas: graph analysis, big data, machine learning, visualization
This study introduces a disaster impact estimation prediction (DIEP) that takes social media self-reports and outputs impact probabilities for FEMA defined classes. The DIEP model is a natural language processing (NLP) model trained and tested on data published by OpenFEMA at zip code level.
Research Areas: Social media in disasters, texas tropical storms, rapid damage prediction, volunteered geographic information (VGI), natural language processing
This paper reviews AI and CV applications in PLF and presents three data-efficient case studies conducted at a research feedlot at Texas A&M University. The assessment of AI model predictability was conducted with precision, recall, F1-score, and adjusted rand index (ARI).
Research Areas: Data-efficient AI, Computer Vision, Individual Animal Identification, Artificial Intelligence, Precision Livestock Farming
Current Collaborators
- Seung Won Yoon, Education and Human Development
- Namgyun Kim, Construction Science
- Liangfei Ye, Sociology
- Michael Fox, ITEC
- Walter R. Magnussen, ITEC
- Patrick Newman, ITEC
- David Retchless, Galveston
- Luis Tedeschi, Animal Science
- Karun Kaniyamattam, Animal Science
- Daniel Paredes-Sabja, Biology
- Zhe Zhang, Geography
- David Florence, TTI
- Zhiwen Fan, Electrical & Computer Engineering
- Cheng Zhang, Computer Science
Past Collaborators
- Tim Lomax, Texas A&M Transportation Institute
- Byron Prestridge, Transportation Services
- Yasuko Sakurai, Utilities & Energy Services
- Ron Steedly, Transportation Service
- Shawn Turner, TTI
- Les Williams, Utilities and Energy Services
- Vivek Gupta, TTI
- Jack Baldauf, Division of Research & Oceanography
- Amir Behzadan, Construction Science
- Costas Georghiades, Department of Electrical & Computer Engineering
- Dan Goldberg, Department of Geography
- Bruce Herbert, University Libraries & Geology/Geophysics
- Tracy Hammond, IEEI & CSE
Publications
- Reddy, J., Shah, N. P., Pi, Y., Lee, J., & Kim, N. (2026). Vehicle speed estimation using a roadside camera and 3D scene reconstruction for work zone intrusion prediction. In Proceedings of the Construction Research Congress 2026. American Society of Civil Engineers.
- Adiga, A., Pi, Y., Choi, K., Chai, D. S., Jung, W., Duffield, N., & Kim, N. (2026). AI-assisted learning for improving construction plan reading skills. Proceedings of the International Conference on Computing in Civil Engineering (i3CE 2026). American Society of Civil Engineers.
- Adekunle, A., Adiga, A., Pi, Y., & Kaniyamattam, K. (2026). A hybrid computer vision pipeline for precision livestock systems: Integrating YOLOv8n and GroundedSAM for real-time cattle re-identification. AI in Agriculture Conference 2026, NCSU.
Kaniyamattam, K., Adekule, A., Vettil, V. K., Rejimon, S. P., Pi, Y., Tao, J., … & Tedeschi, L. O. (2025). 400 Design and development of artificial intelligence-driven decision support systems for sustainable livestock systems in the United States. Journal of Animal Science, 103(Supplement_3), 109–110.
- Cheng, Chih-Shen, Yalong Pi, Tim Lomax, Nick Duffield, and Amir H. Behzadan. 2024. “Pedestrian Phone-Related Distracted Behavior Classification in Front-Facing Vehicle Cameras for Road User Safety.” Computing in Civil Engineering 2023. American Society of Civil Engineers. https://doi.org/10.1061/9780784485248.050.
- Mendes, Egleu D., Jack Wooley, Yalong Pi, Jian Tao, and Luis O. Tedeschi. “419 Automated Individual Animal Identification and Feeding Bunk Scoring: A Computer Vision Approach for Beef Cattle at Calan Gate Feeding System.” Journal of Animal Science 102, no. Supplement_3 (2024): 2–3. https://doi.org/10.1093/jas/skae234.002.
- Pi, Y, J Tao, E. D. Mendes, and L. O. Tedeschi. “Computer Vision Algorithms to Help Decision-making in Cattle Production.” Animal Frontiers 14, no. 6 (2024): 11–22. https://doi.org/10.1093/af/vfae028.
- Z Zhang, S Ren, X Qian, N Duffield (2024). Learning Flexible Time-windowed Granger Causality Integrating Heterogeneous Interventional Time Series Data. ACM SIGKDD Conference on Knowledge Discovery and Data Mining.
- Z Zhang, D Li, Z Zhang, N Duffield (2024). Mining Spatiotemporal Mobility Patterns Using Improved Deep Time Series Clustering. ISPRS International Journal of Geo-Information 13 (11), 374.
- Farinoush Sharifi, Mark Burris, Luca Quadrifoglio, Nick Duffield, Xiaodan Xu, Alexander Meitiv, Yanzhi Ann Xu (2024). Electric truck feasibility for port drayage operations: Port Houston case study. Transportation Research Part D: Transport and Environment, Volume 135, 104355. https://doi.org/10.1016/j.trd.2024.104355
- Ziyi Zhang, Shaogang Ren, Xiaoning Qian, and Nick Duffield (2024). Towards Invariant Time Series Forecasting in Smart Cities. In Companion Proceedings of the ACM Web Conference 2024 (WWW ’24 Companion), May 13–17, 2024, Singapore, Singapore. ACM, New York, NY, USA. https://doi.org/10.1145/3589335.3651897
- Xinyue Ye, Shoujia Li, Ge Gao, David Retchless, Zhenhang Cai, Galen Newman, Jiaxin Du, Nick Duffield. 3D visualization of hurricane storm surge impact on urban infrastructure. Urban Info 3, 9 (2024). https://doi.org/10.1007/s44212-024-00043-x
- Chakraborty, Snehashis, Hernan Santos Barrera, Nick Duffield, Michael Fox, Walter Magnussen, Patrick Newman, Anjuli Renold, Jian Tao, and Valerie Vetrone (2024). “Creating a Shared Public Safety Radio Dataset.” 2024 IEEE World Forum on Public Safety Technology (WF-PST). To appear.
- Raghukumar, H. et al. “Smart Communities, Smart Responders – Artificial Intelligence for Internet of Things Competition: A Case Study in Organizing a Public Safety Innovation Challenge.” 2024 IEEE World Forum on Public Safety Technology (WF-PST). To appear.
- Mendes, Egleu D., Yalong Pi, Jian Tao, and Luis O. Tedeschi. “110 Evaluation of Computer Vision to Analyze Beef Cattle Feeding Behavior.” Journal of Animal Science 101, no. Supplement_3 (2023): 2–3. https://doi.org/10.1093/jas/skad281.003.
- Pi, Yalong, Egleu Mendes, Luis Tedeschi, Jian Tao, Siddhanth Reddy, and Nick Duffield. “Using Computer Vision Pixel Segmentation to Detect Beef Cattle Feeding Behavior.” 2023. Submitted.
- Tao, Jian, Egleu D. Mendes, Yalong Pi, Andrew Cassity, R. R. Male, K. Kaniyamattam, et al. “Hands-On III: Building Digital Twins for Precision Livestock Farming: Data Analytics and Big Data Challenges.” Journal of Animal Science 101, no. Supplement_3 (2023): 75–76. https://doi.org/10.1093/jas/skad281.092
- Li, Diya, Zhe Zhang, Bahareh Alizadeh, Ziyi Zhang, Nick Duffield, Michelle A. Meyer, Courtney M. Thompson, Huilin Gao, and Amir H. Behzadan. “A reinforcement learning-based routing algorithm for large street networks.” International Journal of Geographical Information Science (2023). DOI: 10.1080/13658816.2023.2279975.
- Zhang, Ziyi, Diya Li, Zhenlei Song, Nick Duffield, and Zhe Zhang. “Location-Aware Social Network Recommendation via Temporal Graph Networks.” Proceedings of the 7th ACM SIGSPATIAL Workshop on Location-based Recommendations, Geosocial Networks and Geoadvertising (LocalRec ’23), 2023, pp. 58–61. DOI: 10.1145/3615896.3628342.
- Sharifi, Farinoush, Mark Burris, Luca Quadrifoglio, Nick Duffield, Xiaodan Xu, and Alexander Meitiv. “Heavy-duty Battery Electric Trucks for Port Drayage Operations: A Feasibility Study.” Transportation Research Board 2023 Annual Meeting, 2023.
- Chih-Shen, C., Yalong Pi, Tim L., Nick D., and Amir H. B. “Pedestrian phone-related distracted behavior classification in front-facing vehicle cameras for road user safety.” I3CE 2023: 2023 ASCE International Conference on Computing in Civil Engineering. https://ascelibrary.org/doi/10.1061/9780784485248.050
- Pi, Yalong, Nick Duffield, Amir Behzadan, and Tim Lomax. “Lane-specific speed analysis in urban work zones with computer vision.” Traffic Injury Prevention 24, no. 3 (2023): 242–250. DOI: 10.1080/15389588.2023.2173522.
- Pi, Yalong, Xinyue Ye, and Nick Duffield. “Rapid Disaster Damage Estimation Based On Geo-referenced Social Media Data.” AGU Fall Meeting Abstracts, vol. 2022, December 2022, pp. NH14A-03.
- Duffield, Nick. “Inspiring Data Science Research through Collaboration with University Operations.” Keynote Talk, ACM Sigmetrics / IFIP Performance Conference, 2022.
- Pi, Yalong, Xinyue Ye, Nick Duffield, and Microsoft AI for Humanitarian Action Group. “Rapid Damage Estimation of Texas Winter Storm Uri from Social Media Using Deep Neural Networks.” Urban Sci. 6, no. 3 (2022): 62. https://www.mdpi.com/2413-8851/6/3/62
- Pi, Y., T. Lomax, N. Duffield, and A. Behzadan. “Visual Recognition for Traffic Information Retrieval and Analysis in Major Events Using Convolutional Neural Networks.” Comp. Urban Sci. 2, no. 2 (2022). https://link.springer.com/article/10.1007/s43762-021-00031-w
- Lin, B., L. Zou, N. Duffield, A. Mostafavi, H. Cai, B. Zhou, et al. “Revealing the linguistic and geographical disparities of public awareness to Covid-19 outbreak through social media.” International Journal of Digital Earth 15, no. 1 (2022): 868–889. https://doi.org/10.1080/17538947.2022.2070677
Zhang, Ziyi, Diya Li, Zhe Zhang, and N. Duffield. “A time-series clustering algorithm for analyzing the changes of mobility pattern caused by COVID-19.” HANIMOB ’21: Proceedings of the 1st ACM SIGSPATIAL International Workshop on Animal Movement Ecology and Human Mobility, November 2021. https://dl.acm.org/doi/10.1145/3486637.3489489
Funding
[TAMIDS 2026-2027] Texas A&M Institute of Data Science. Generative AI Learning Assistance for Improving Students’ Construction Plan Reading Skills. PI: Namgyun Kim. Co-PI: Yalong Pi. Total $2,000 + 5,000 OpenAI credits.
[TAMU 2026/06-2027/06] Texas A&M College of Arts & Sciences Environment and Sustainability Initiative (ESI). Interpretable AI for Compounding and Cumulative Climate Risks in Cognitive Aging: Mixture and Deep Temporal Models Using HRS/HRS-CDR. PI: Liangfei Ye. Co-PI: Yalong Pi, Jessica A. Bernard. Total $10,000.
[NSF 2025-2028] National Science Foundation. Collaborative Research: Empowering STEM Education by Creating Benchmarking Models for Generative Artificial Intelligence Learning Assistants. PI: Namgyun Kim. Co-PIs: Yalong Pi, Dae Seok Chai, Nicholas Duffield, Kunhee Choi. Total $225,043.
[TAMIDS 2025-2027] Texas A&M Institute of Data Science. An Integrated Digital Twin and Computer Vision Model for a Proactive Work Zone Intrusion Warning System. PI: Namgyun Kim. Co-PI: Yalong Pi. Total $20,000.
[AgriLife 2025-2026] Texas A&M AgriLife Research. A hybrid framework for combining precision livestock technology and decision modeling for disease prediction in beef systems. PI: Karun Kaniyamattam. Co-PIs: Luis Tedeschi, Renaldo Cooke, Gizem Levent, Yalong Pi. Total $47,459.
[NVIDIA 2025-2026] NVIDIA. Advancing Precision Livestock Management: NVIDIA-Powered Computer Vision System for Bovine Respiratory Disease Detection. PI: Karun Kaniyamattam. Co-PI: Yalong Pi, Jian Tao. Total $ 32,000 computing hours on Saturn Cloud.
[TAMIDS 2024-2026] Texas A&M Institute of Data Science. Developing a Computer Vision Data Pipeline for Bovine Respiratory Disease Prediction in Beef Cattle. PI: Karun Kaniyamattam. Co-PI: Yalong Pi. Total $ 40,000.
[USDA-NIFA 2023-2026] United States Department of Agriculture National Institute of Food and Agriculture. Harnessing Precision Livestock Farming to Support Smart Agriculture for Sustainable Beef Cattle Production. PI: Luis Tedeschi. Co-PIs: Jian Tao, Nicholas Duffield, Yalong Pi, Karun Kaniyamattam. Total $ 106,766.
[TOGCE 2023-2025] Texas One Gulf Center of Excellence. Towards Targeted Risk Mitigation: Community Engaged, Fast Impact Estimation of Extreme Weather using Big Social and Climate Data. PI: David Retchless, Co-PIs: Xinyue Ye, Yalong Pi, Nicholas Duffield, and Wesley Highfield. Total $532,988.
[USDA 2023] Harnessing Precision Livestock Farming to Support Smart Agriculture for Sustainable Beef Cattle Production, USDA-NIFA, PI L. Tedeschi, Co-PIs N. Duffield, Y. PI, J. Tao, 09/01/2023-08/31/2024, $103,563
[NIST 2022] A Digital-Twin Enabled Testbed for Public Safety Communication Technologies, National Institute of Standards and Technology, PI. J. Tao, Co-PIs N. Duffield, W. Magnussen, A. Thomas, N. Reddy, 06/01/2022-05/31/2024, $1,200,000.
[NIST 2021] Creating a Shared Public Safety Radio Data Set for Sharing and Analysis, National Institute of Standards and Technology, 09/01/2021-08/31/2023, PI W. Magnussen, Co-I N. Duffield, J. Tao. Total $904,806 / Duffield $309,566
[MS 2021] Rapid damage prediction from social media using historical big data and deep learning. Microsoft AI for Humanitarian Action Program, PI Y. Pi, Co-PIs N. Duffield, X. Ye, $150,000 Azure Computing Credits