Chris Comiskey, Ph.D. Email & Phone Number

Amsterdam, North Holland, Netherlands

* Customer Service Call Forecasting with SARIMAX models* MAB for Personalized Marketing Actions* Reinforcement Learning for Optimal Service Actions and Call Routing

Amersfoort, Utrecht, Netherlands

* Time series forecasts of Radio Network usage KPIs, across 15K sectors* Anomaly Detection for Network security* Network reconstruction process analysis* Network fault detection algorithm design

Rotterdam Area, Netherlands

* Fraud detection with Random Cut Forest anomaly detection* AWS Sagemaker, Bitbucket, Dataiku, Python OOP and development* Forecasting (cashflow, churn) w/ custom time series models * Optimized risk assessment modelling

Amsterdam Area, Netherlands

* Senior Data Scientist at adidas, Digital Advanced Analytics, EU eCommerce* Delivered probabilistic demand forecasts informing €100m in intake quantity (GAS models)* Improved purchasing strategy with custom Optimal Overbuy methodology* Developed Hierarchical Sampling technique for new-article demand forecasting

Amsterdam Area, Netherlands

• Senior Data Scientist at adidas, Digital Advanced Analytics • Preseason demand forecasts informing €70m in EU eCommerce intake quantity• Counterfactual Analysis of iROAS in Geo experiment for Traffic Generation Analytics• Article classification with k-means, agglomerative, and DBSCAN algorithms• LIME for machine learning model interpretations

Chicago, Illinois

• Predicted mechanical failure probabilities with Survival Analysis decision trees for Ingersol Rand• Predict Consumer Price Index with ARIMA models, for Exos Financial Services• Develop decision tree/random forest implementation R package for Prudential Financial• Scrape web data for Exos Financial• Literature review of Survival Analysis decision tree methodologies for Ingersol Rand• Develop R SDK components for software product REST API• Git, GitHub, Bash, MySQL, Python, HTML, Docker• R Markdown reports, LaTeX Beamer presentations

Corvalis, Or

• Developed variable-resolution heat maps for spatial data• Developed R package varyres for implementing variable-resolution heat maps• Developed interactive heat map confidence intervals for spatial estimators• Developed R package mapapp to create interactive heat map confidenceintervals with RStudio’s Shiny• Estimated “Effective Sample Size” for hydrologic AR(1) time series• Modeled spatially correlated Bernoulli random variables with Bayesian hierarchicallogistic regression (with GRF) model

Corvallis, Oregon

• Graduate Teaching Assistant, OSU Statistics Department• Consulting, OSU Student Consulting Services• Course development, OSU Data Analytics M.S. • ST 516 – Foundations of Data Analytics • ST 517 – Data Analytics I

Doctor Of Philosophy (Phd), Statistics

"Take Me Out to (Analyze) the Ballgame: Visualization and Analysis Techniques for Big Spatial Data"

Master Of Science (Ms), Statistics

Effective sample size estimation for trend detection in hydrologic time series.

Bachelor Of Science (B.S.), Mathematics