Gorkem Turgut (G.T.) Ozer, MBA, MSc, PhD

Picture recreated by Nano Banana – Click for the original

In the last 15+ years, I have used statistical modeling and machine learning to solve causal (and predictive) problems, managed data science teams, advised at the executive level on business analytics and data centricity, and mentored tech startups on digital platform strategy. My "data science" expertise is centered on causal inference using statistical, machine learning, and Bayesian methods, and in exploring what-if scenarios or counterfactuals via agent-based models and ongoing work involves quasi-experimental problem solving.

Previously founded two companies, and held corporate roles driving IT strategy, digital transformation, and data science initiatives.

For any updates and posts on data science, causal models, and AI, visit ozer.gt/log (typically cross-posted at linkedin.com/in/gtozer). For the slide decks, including Mind the AI Gap: Understanding vs. Knowing, AI in 64 Pictures, and Data Centricity, visit ozer.gt/log/talks.

Research

Areas: Strategic and societal implications of digital platforms and technology; Role of digitization, algorithms, and AI (LLM) agents; Methodological advancements in Information Systems & Economics

Published:

"The Sources of Researcher Variation in Economics." Journal of Economic Literature, 0(0), 2026 (Forthcoming). → Multi-author collaboration

"Let it Ride! An Empirical Investigation of Problem Gambling and the Implications of Legalized Online Sports Betting." Information Systems Research (An INFORMS Journal), 0(0), 2026. doi.org/10.1287/isre.2024.1526

"On Influence in Online Creative Craft Communities and the Relevance of Virtuosity and Community Engagement." Journal of the Association for Information Systems, 27(3), 693-718, 2026. doi.org/10.17705/1jais.00982

"Noisebnb: An Empirical Analysis of Home-Sharing Platforms and Residential Noise Complaints." Information Systems Research, 35(4): 1824-1847, 2024. pubsonline.informs.org/doi/10.1287/isre.2022.0070

"Digital Multisided Platforms and Women's Health: An Empirical Analysis of Peer-to-Peer Lending and Abortion Rates." Information Systems Research, 34(1):223-252, 2022. pubsonline.informs.org/doi/10.1287/isre.2022.1126
→ INFORMS ISS Bapna-Ghose Social Justice Best Paper Award

"Combining Stock-and-flow, Agent-based, and Social Network Methods to Model Team Performance." System Dynamics Review, 34: 527-574, 2020. onlinelibrary.wiley.com/doi/abs/10.1002/sdr.1613

Ongoing:

"Are LLM Agents Behavioral Consumers When Tasked with Shopping? Evidence from a Pricing & Merchandising Experiment on Amazon Product Cards"

"First, Do No Harm: A Dual-Process Model of Framing and Source Provision in Trust in Large Language Models"

"Can Algorithms Title What They Did Not Create? Human-Algorithm Complementarity in Photography as Creative Craft"

"To Ensemble or Not: Empirical Analysis of Incremental Performance Gains in Predictive Hazards Models"

"The Voice to the People in Times of Trouble: The Impact of the Floyd Video on Hip-hop Music Listenership in the U.S."

"Organizational Learning in the Context of Multisided Digital Platforms: A Multi-method Simulation Study"

Teaching

Areas: Data science, causal and predictive modeling, AI for business; LLM agents and agentic workflows; Tech, digital platforms, and cloud

Courses I've developed and taught at the University of Texas at Austin, University of Maryland, College Park, and University of New Hampshire:

Big Data and Artificial Intelligence: Strategy & Analytics ozer.gt/bigdata
Analysis of big data using modern data science tools: Data Centricity and Business Value of AI, MLOps, Causal AI and Inference in Big Data, Reinforcement Learning, and Generative Models. Uses both R & Python supported by AI coding assistants in VS Code and Cursor.

AI Tools and Applications ozer.gt/aitools
Docker; Git, GitHub, and Hugging Face; LLMs, AI Agents, and Agentic Workflows for Data Science (using VS Code & Cursor); AI automation.

Predictive Analytics and Modeling ozer.gt/predict
Covers modern predictive analytics methods: Logistic, Probit, and Poisson Regression; Decision Trees and Ensemble Methods (Random Forests, XGBoost, LightGBM, CatBoost); Lasso and Ridge Regression; Deep Learning and Neural Networks. Uses R with VS Code and Cursor.

Predictive Analytics for Business ozer.gt/data
Selected statistical and machine learning methods along with strategy and business implications. Uses R as the primary analytical tool, a first for the MBA program, and encourages the use of AI coding assistants.

Managing Technology: Strategy, Software, and Data ozer.gt/301
Three modules: Strategy in technology ecosystems; Software is eating the world; Data centricity: When strategy meets software. Addresses tech-enabled business models, IS/IT fundamentals (databases, agile IT project management, cybersecurity), and the impact of information technology on today's businesses and management.

Introduction to IT Management ozer.gt/introtoit
Robust adaptive strategies, competitive advantage in the digital age, multisided platforms and network effects, digital transformation, database management, cloud computing, cybersecurity, software: DevOps & agile.

Applied Cloud Computing using AWS ozer.gt/aws (Student-initiated)
Prep for Amazon Web Services Solutions Architect certification by covering tools and tech needed to architect and deploy robust and secure applications, core architectural design principles, and their applications.

Picture recreated by Nano Banana – Click for the original

Project Highlights

Data Centricity Lab — datacentricity.org

Data Centricity Lab builds models and agentic workflows that turn research into decision-ready insights for business. The underlying portfolio includes causal research on the price elasticity of LLM agents in retail, a technical business blog, and proof-of-concept apps across AI for data science, retail, and the future of work, built on custom agent frameworks. An example prototype is at findcredible.com; the technical blog is at dataduets.com.

Causal Book — causalbook.com

This book offers a curated set of design patterns for causal inference in data science, with modern applications of each pattern across three approaches: Statistics (traditional frequentist methods), Machine Learning, and Bayesian. Each design pattern is supported by business cases that use realistic or real data for causal inference. The three approaches are compared using the latest tools on the same data and model in R and Python. Work-in-progress.

Pedagogical Innovations

In-Class Hackathons: Predictive Analytics (modeling hackathon) and Reinforcement Learning (multi-armed bandits), both in competition format.

Kaggle Competitions: Causal inference and predictive modeling using rich panel data: 10+ years of data with 15 million observations on 115 variables.

AI-assisted Production: Students develop a machine learning model and deploy it to Hugging Face for live access as a tangible portfolio piece.

DataScience.Day: Speaker series focused on the application of course concepts to business problems and providing networking opportunities.