Davut Ayan

Davut Ayan

Senior Data Scientist

Starcom (Publicis Groupe)

I build rigorous, scalable analytics that turn complex data and ambiguous business questions into practical decisions.

I am a Senior Data Scientist at Starcom (Publicis Groupe) in Chicago and a PhD-trained economist with 7+ years of experience across marketing, media, public policy, and risk analysis. My work combines machine learning, causal inference, experimentation, generative AI, and large-scale data workflows using Python, R, SQL, PySpark, Databricks, Snowflake, and Azure.

Download my resumé .

Interests
  • Causal Inference
  • Machine Learning
  • Generative AI
Education

  • PhD in Economics, 2016

    University of Kansas

  • MA in Economics, 2013

    University of Kansas

  • BS in Industrial and Systems Engineering, 2007

    National Defense University

Core Expertise

Machine Learning & AI

Predictive modeling, supervised and unsupervised learning, model validation, NLP, and generative-AI-assisted analytics

Causal Inference & Experimentation

Quasi-experimental designs, panel methods, A/B testing, measurement strategy, diagnostics, and interpretation

Marketing & Audience Analytics

Customer segmentation, lifetime value, market sizing, campaign measurement, and behavioral analysis

Programming & Modeling

Python, R, SQL, PySpark, statistical modeling, reusable analytical code, and visualization

Data Platforms

Databricks, Snowflake, Azure, cloud analytics, scalable pipelines, and data-quality workflows

Analytics Products

Automated reporting, Streamlit and Shiny applications, decision-support tools, and stakeholder-ready insights

Selected Work

*
All Causal Inference Machine Learning Gen AI Other
Next Word Prediction
This project is a R Shiny Web Application that uses Stupid Back-off N-gram models, a Large Language Model, to predict next words given input text.
Last updated on Jun 28, 2024 6 min read LLM
Repo Slides
Web Scraping using rvest package
This project showcases how rvest package in R can be utilized to scrape data from a website.
Last updated on Jun 28, 2024 5 min read R

Experience

 
 
 
 
 
Senior Data Scientist
Starcom (Publicis Groupe)
October 2025 – Present Chicago, IL (Hybrid)
  • Lead audience analytics, customer segmentation, transaction-based measurement, and automated reporting workstreams.
  • Build scalable SQL, Python, PySpark, and Databricks workflows for model-ready features, audience development, and decision support.
  • Partner with product, engineering, governance, analytics, and business teams to align definitions, resolve data-quality issues, and deliver reusable outputs.
 
 
 
 
 
Data Scientist, Associate Director
Horizon Media
March 2022 – October 2025 New York, NY
  • Developed predictive, segmentation, market-sizing, campaign-measurement, and customer analytics workflows.
  • Built reusable tools with Python, R, SQL, Snowflake, Streamlit, and cloud analytics platforms.
  • Applied statistical modeling and generative-AI-assisted workflows to translate audience data into marketing recommendations.
 
 
 
 
 
Postdoctoral Researcher
University of Kansas
February 2021 – March 2022 Lawrence, KS
  • Conducted applied policy research using regression discontinuity, synthetic control, and panel-data methods.
  • Developed reproducible workflows and data dictionaries for administrative and public datasets.
 
 
 
 
 
Lead Economist, Lecturer
National Defense University
August 2016 – February 2021 Turkey
  • Taught economics and contributed to curriculum development and applied instruction.
  • Conducted risk analysis and supported disaster and emergency-management planning and exercises.

Learning Library

I maintain a growing collection of books and technical notes for structured study, practical reference, and data-science interview preparation. The materials cover programming, statistics, causal inference, machine learning, analytics, and related tools.

Explore the Learning Library →

Contact