Portfolio & Leadership

Analytics in Action

Sports Analytics

Performance Metric Validation Study: Challenged vendor-provided performance analytics that suggested a critical skill area required significantly more practice investment. Through comparative analysis across multiple data sources and creation of synthetic benchmarking datasets, identified potential measurement inconsistencies and provided context that reframed the team’s relative performance position.

Key findings revealed that data collection practices were introducing systematic bias, and that apparent weaknesses were partially offset by exceptional performance in complementary skill areas. Recommendations led to improved measurement protocols and more accurate self-assessment capabilities for athletes, while enabling more strategic resource allocation for coaching staff.

Methods: Comparative benchmarking, synthetic data modeling, distributional analysis, relative performance metrics

Higher Education

Enrollment Yield Prediction & Strategic Recruiting Framework: Leading predictive analytics initiatives for a major public university managing 50,000+ annual applicants, addressing critical gaps in enrollment forecasting and recruiting strategy. Developed rigorous, evidence-based models that identify approximately 80% of likely depositors four months before the May 1st deadline, enabling strategic resource allocation across admissions counseling teams.

The modeling framework integrates economic theory, demographic analysis, and machine learning to overcome limitations of previous vendor solutions and ad hoc approaches. By synthesizing diverse data sources and methodological strengths across disciplines, created actionable intelligence that transforms how the institution targets and supports prospective students.

Complementary work on geographic recruiting optimization identified underperforming communities by benchmarking demographic profiles against high-yield peer regions across Michigan and nationwide. This analysis revealed untapped recruiting opportunities in communities with similar composition to successful markets, enabling data-driven expansion of recruiting territories.

Models scheduled for operational deployment in January 2026, with anticipated improvements in yield prediction accuracy and counselor efficiency.

Methods: Big data, predictive analytics, demographic clustering, comparative benchmarking, longitudinal modeling, geographic market analysis

Building Analytics Excellence

Innovation & Collaboration

Analytics Infrastructure

Open-source tools and frameworks are a key component of exeResearch’s design philosophy. Combining mainstream applications and methods with custom functions to solve unique problems reduces costs for our clients. The following are several tools to solve real-world analytics challenges:

Higher Education Analytics

  • HovedTraek - The HovedTraek package is focused on feature engineering for enrollment predictive modeling. The package transforms applicant data into temporal behavior derivatives used in deposit and cancellation likelihood prediction models. It is application-platform agnostic and powers the enrollment predictions for 50,000+ applicants for the Fall 2026 incoming class. .:|:. In-house application - approach transferable to similar use cases.
  • FOLI - The Feature-Observation Landscape Index (FOLI) package leverages the concept of entities (communities, people, and materials to name a few) with similar features should behave in a same manner. FOLI identifies regions following and diverging from this principle. Developed to serve the demands of MSU’s Office of Admissions, FOLI now possess prospect (screening) and target (recommendation) identification capabilities as it enters its second year of use. .:|:. In-house application - approach transferable to similar use cases.
  • theHUB - The Haabefuld Utility Box (theHUB), is a comprehensive analytics toolkit for higher education institutions, featuring functions, scripts, and worked examples for student success analysis. Initially developed for Michigan State University, it is a collection of analysis tools composed of functions, scripts, electronic notebooks, and worked examples. theHUB was designed to share these tools with the MSU community to improve the analyses of student outcomes, student success, and survey responses.

Sports Analytics

  • RaeveStoberi - The RaeveStoberi package is our in-house application for harvesting, extracting, and cleaning sports-related data. RaeveStoberi currently serves our need for constructing player, team, game, play-by-play, and league datasets for the NHL, WNBA, Unrivaled Basketball League, and various NCAA sports. .:|:. In-house application - constantly expanding to support new sports projects.

Data Visualization

  • MSUthemes - The MSUthemes package provides colour palettes and themes for Michigan State University (MSU) and comprehensive colour support for all Big Ten Conference institutions. The package includes MSU-specific palettes (sequential, diverging, and qualitative) designed to align with MSU's branding guidelines, uses the Metropolis font (MSU's chosen font) in the construction of the plots, as well as the primary and secondary colour palettes for all 18 Big Ten institutions, making it ideal for multi-institutional comparisons and collaborative research visualizations using ggplot2.