4th Down Decisions: NFL Analytics Capstone

This project analyzes 4th down decision-making in the NFL from 2018 to 2023, comparing actual coaching decisions to win probability models.

Overview

Using real NFL play-by-play data and tools like SQL, R, and Tableau, I identified trends, coaching patterns, and potential inefficiencies in 4th down strategies.

Key Objectives

• Determine aggressiveness of NFL teams on 4th down from 2018 to 2023
• Measure alignment with analytics-based win probability models
• Visualize tendencies by coach, distance, and field position

Tools Used

• SQL – Data wrangling
• R – Statistical modeling
• Tableau – Interactive dashboards
• GitHub – Version control
• NFLfastR – Data source

Key Files

• nfl_4th Down Dataset.csv: Cleaned dataset
• Tableau_Dashboard.png: Interactive Visual
• 4th Down WP.Rmd: Predictive modeling notebook in R
• 4th Down WP.ipynb: Predictive Modeling Notebook in Python
• Capstone Summary Slide Deck.pdf: Slide summary

Highlights

• Identified a +2.8% average WP swing for aggressive teams like the Ravens
• Found that 37% of punts between the opponent’s 40–50 were suboptimal
• Built dashboards for quick comparison of coach tendencies

Next Steps

• Expand to include overtime and playoff scenarios
• Incorporate defensive strength into decision-making models
• Automate data refresh using scheduled R scripts or Python pipelines

---

Let’s connect if you’re interested in data storytelling, sports analytics, or business intelligence!