Part 7: Capstone Projects

Introduction

These capstone projects integrate concepts from across the entire textbook, challenging you to build complete analytical systems that could be used in professional settings. Each project includes specifications, milestones, evaluation criteria, and extension opportunities.

Choose projects based on your interests and career goals. Each can be completed individually or adapted for team collaboration.


Project 1: Complete Game Prediction System

Overview

Build an end-to-end NFL game prediction system that forecasts point spreads and totals for every game of the season.

Learning Objectives

  • Integrate power ratings, efficiency metrics, and contextual adjustments
  • Build and maintain a prediction database
  • Track and evaluate model performance
  • Implement automated weekly updates

Requirements

Core Components: 1. Team power rating system with weekly updates 2. Home field advantage adjustments (venue, schedule, travel) 3. Weather impact modeling 4. Injury adjustment framework 5. Point spread and total predictions 6. Performance tracking and calibration

Technical Requirements: - Python implementation with pandas/numpy - Database storage (SQLite or PostgreSQL) - Automated data collection pipeline - Visualization dashboard (matplotlib/plotly)

Milestones

Week 1-2: Foundation - Set up data pipeline for historical games (2018-present) - Implement basic Elo rating system - Calculate initial power ratings

Week 3-4: Adjustments - Add home field advantage (static and dynamic) - Implement schedule factors (bye weeks, rest) - Add weather adjustments

Week 5-6: Injuries and Context - Build injury impact model by position - Add playoff/rivalry adjustments - Integrate market data for validation

Week 7-8: Automation and Evaluation - Build weekly update pipeline - Create performance tracking dashboard - Document calibration and accuracy metrics

Deliverables

  1. Working prediction system with documentation
  2. Historical backtest results (3+ seasons)
  3. Weekly prediction export mechanism
  4. Performance report with accuracy metrics

Evaluation Criteria

Criterion Weight Description
Technical Implementation 30% Code quality, architecture
Prediction Accuracy 25% ATS performance, MAE
Integration 20% Successful component integration
Documentation 15% Clear explanation of methodology
Extensions 10% Additional features beyond requirements

Extensions

  • Add player prop predictions
  • Implement Monte Carlo simulation for uncertainty
  • Build web interface for predictions
  • Add live game probability updates

Project 2: Fantasy Football Analytics Platform

Overview

Create a comprehensive fantasy football analysis platform that provides projections, rankings, and decision support.

Learning Objectives

  • Apply VORP and positional scarcity concepts
  • Build player projection models
  • Implement DFS optimization
  • Create actionable decision tools

Requirements

Core Components: 1. Player projection engine 2. VORP-based ranking system 3. Draft optimization tool 4. Weekly start/sit recommendations 5. DFS lineup optimizer 6. Waiver wire analyzer

Technical Requirements: - Python with pandas/scipy - Linear programming for DFS - Web interface (Flask/Streamlit optional) - Data integration with fantasy platform APIs

Milestones

Week 1-2: Projections - Build player projection model - Implement regression to mean - Create efficiency and volume forecasts

Week 3-4: Rankings - Calculate replacement levels - Implement VORP rankings - Create position scarcity analysis

Week 5-6: Decision Tools - Build DFS optimizer - Create start/sit recommender - Implement waiver analyzer

Week 7-8: Integration - Build unified platform - Create export/visualization features - Test with live data

Deliverables

  1. Projection database with weekly updates
  2. VORP rankings with positional analysis
  3. DFS optimizer with multiple contest types
  4. Documentation of methodology

Evaluation Criteria

Criterion Weight Description
Projection Accuracy 25% Compare to expert consensus
VORP Implementation 25% Correct value calculations
Optimization Quality 20% Valid lineups, value generation
Usability 20% Clear outputs, actionable
Documentation 10% Methodology explanation

Extensions

  • Dynasty/keeper valuations
  • Trade analyzer
  • Auction value calculator
  • Best ball optimization

Project 3: NFL Draft Prospect Evaluator

Overview

Build a comprehensive prospect evaluation system that ranks draft prospects and projects NFL success.

Learning Objectives

  • Apply production normalization techniques
  • Integrate combine and athletic testing
  • Build position-specific models
  • Create comparable player identification

Requirements

Core Components: 1. College production database 2. Conference adjustment system 3. Combine analysis module 4. Position-specific evaluation models 5. Draft value calculator 6. Historical validation framework

Technical Requirements: - Python with scikit-learn for modeling - Historical prospect database (2015+) - Visualization for profiles - Exportable reports

Milestones

Week 1-2: Data Foundation - Build college stats database - Implement conference adjustments - Create production metrics (YPRR, Dominator, etc.)

Week 3-4: Athletic Profiles - Process combine data - Calculate Speed Score, RAS equivalents - Build athletic comparison system

Week 5-6: Models - Build QB, WR, RB projection models - Implement comparable identification - Create draft range predictions

Week 7-8: Validation - Backtest against historical classes - Calculate accuracy metrics - Document strengths/weaknesses

Deliverables

  1. Prospect database with comprehensive metrics
  2. Position-specific evaluation tools
  3. Draft board generation system
  4. Validation report with historical accuracy

Evaluation Criteria

Criterion Weight Description
Data Quality 20% Complete, accurate data
Model Accuracy 30% Historical prediction performance
Feature Engineering 20% Creative, useful metrics
Usability 15% Practical output format
Documentation 15% Methodology transparency

Extensions

  • Add film-based metrics (PFF integration)
  • Build trade value calculator
  • Create team-specific fit analysis
  • Implement machine learning models

Project 4: Advanced EPA Analysis System

Overview

Build a complete Expected Points Added (EPA) analysis system using play-by-play data.

Learning Objectives

  • Calculate and interpret EPA from first principles
  • Build success rate and efficiency metrics
  • Create player and team evaluation systems
  • Implement advanced analytics visualizations

Requirements

Core Components: 1. EPA calculator from play-by-play 2. Expected Points model 3. Success rate framework 4. Player EPA attribution 5. Team efficiency dashboards 6. Situation-specific analysis

Technical Requirements: - nflfastR or equivalent data source - Python/R for analysis - Interactive visualizations - Database for calculated metrics

Milestones

Week 1-2: EPA Foundation - Load play-by-play data - Calculate expected points by field position - Implement basic EPA

Week 3-4: Success Rate - Define success thresholds by down/distance - Calculate success rate by team/player - Build explosive play metrics

Week 5-6: Player Analysis - Attribute EPA to players (QB, RB, WR) - Create CPOE calculations - Build receiver efficiency metrics

Week 7-8: Visualization - Create team dashboards - Build player comparison tools - Implement situation analysis

Deliverables

  1. EPA calculation engine
  2. Team and player efficiency reports
  3. Interactive visualization dashboard
  4. Methodology documentation

Evaluation Criteria

Criterion Weight Description
Technical Accuracy 30% Correct EPA calculations
Analysis Depth 25% Insights beyond basic metrics
Visualization 20% Clear, informative charts
Usability 15% Accessible outputs
Documentation 10% Clear methodology

Extensions

  • Add win probability model
  • Implement clutch metrics
  • Create coaching evaluation system
  • Build play-call analysis

Project 5: Sports Betting Analysis Tool

Overview

Create a comprehensive betting market analysis tool that tracks lines, identifies value, and evaluates performance.

Learning Objectives

  • Work with odds and probability conversions
  • Track market movements
  • Implement CLV analysis
  • Build bankroll management systems

Requirements

Core Components: 1. Odds database and tracking 2. Implied probability calculator 3. Model vs market comparison 4. CLV tracking system 5. Performance analytics 6. Bankroll optimization

Technical Requirements: - Market data integration - Time-series tracking - Statistical analysis tools - Reporting system

Milestones

Week 1-2: Market Data - Build odds tracking database - Implement probability conversions - Track line movements

Week 3-4: Model Comparison - Integrate prediction model - Calculate edge vs market - Implement CLV tracking

Week 5-6: Performance - Build betting log system - Calculate ROI metrics - Analyze by market/bet type

Week 7-8: Optimization - Implement Kelly Criterion - Create bankroll projections - Build reporting dashboard

Deliverables

  1. Market tracking database
  2. Edge calculation system
  3. Performance analytics dashboard
  4. Bankroll management tool

Evaluation Criteria

Criterion Weight Description
Data Accuracy 25% Correct odds handling
Analysis Quality 25% Meaningful insights
Risk Management 20% Sound bankroll approach
Performance Tracking 20% Complete metrics
Documentation 10% Clear explanations

Extensions

  • Add prop bet analysis
  • Implement arbitrage detection
  • Create market efficiency tests
  • Build automated alerts

Project 6: Comprehensive Team Analytics Report

Overview

Create a season-long analytical report for a specific NFL team, similar to what a team analytics department might produce.

Learning Objectives

  • Apply all analytical frameworks to a single team
  • Create executive-level summaries
  • Build comprehensive evaluation systems
  • Produce professional reports

Requirements

Core Components: 1. Team efficiency profile 2. Personnel analysis 3. Scheme evaluation 4. Schedule/opponent analysis 5. Prediction and simulation 6. Executive summary

Technical Requirements: - Complete data integration - Professional visualization - Written analysis components - Exportable report format

Milestones

Week 1-2: Team Profile - Calculate team power ratings - Build offensive/defensive efficiency analysis - Create personnel valuations

Week 3-4: Deep Analysis - Evaluate coaching decisions - Analyze play-calling tendencies - Study situational performance

Week 5-6: Season Context - Analyze schedule strength - Project remaining games - Simulate playoff scenarios

Week 7-8: Report Creation - Compile executive summary - Create visualizations - Document methodology

Deliverables

  1. Complete team analytics report (20+ pages)
  2. Supporting data and visualizations
  3. Executive summary (2 pages)
  4. Methodology appendix

Evaluation Criteria

Criterion Weight Description
Analysis Depth 30% Comprehensive coverage
Insight Quality 25% Actionable findings
Presentation 20% Professional format
Data Integration 15% Multiple sources combined
Documentation 10% Clear methodology

Extensions

  • Add video/film integration
  • Create interactive dashboard
  • Build real-time updates
  • Produce coach-specific recommendations

General Guidelines

Project Selection

Choose a project that: - Aligns with your career interests - Challenges your current skills appropriately - Can be completed within the timeframe - Provides portfolio-worthy output

Timeline

Most projects are designed for 8 weeks at 10-15 hours/week. Adjust milestones if working on a different schedule.

Collaboration

These projects can be: - Individual (recommended for portfolio building) - Team-based (divide responsibilities clearly) - Mentored (with instructor guidance)

Submission

Include: 1. Working code (GitHub repository recommended) 2. Written documentation 3. Sample outputs/visualizations 4. Reflection on methodology

Grading Rubric (If Academic)

Component Weight
Technical Implementation 35%
Analytical Quality 30%
Documentation 20%
Presentation 15%

Resources

Data Sources: - nflfastR/nflfastpy for play-by-play - Pro Football Reference for historical - ESPN API for current stats - FantasyPros for projections

Tools: - Python (pandas, numpy, scipy, scikit-learn) - R (tidyverse, nflfastR) - Visualization (matplotlib, seaborn, plotly) - Database (SQLite, PostgreSQL)

References: - All chapters from this textbook - Further reading sections - Code examples provided


Conclusion

These capstone projects represent the culmination of your NFL analytics education. Each integrates multiple concepts and challenges you to build complete, professional-quality systems. The skills developed here directly transfer to careers in team analytics, fantasy sports, betting markets, and sports media.

Select a project that excites you, plan your approach carefully, and build something you're proud to show. The best analytics work combines technical skill with creative insight—let these projects showcase both.