This project applies a historical NBA success prediction model to the 2026 NBA Draft class to identify players whose pre-draft profiles suggest they may provide greater long-term value than their draft position alone indicates.
The model is designed to complement traditional scouting by using historical NBA outcomes, college production, advanced efficiency metrics, and physical measurements to evaluate every drafted player.
- Evaluated all 60 players selected in the 2026 NBA Draft.
- Applied a historical NBA success model trained on 382 NBA draft selections (2012–2020).
- Used college production, physical measurements, and shooting efficiency metrics to generate player success scores.
- Produced player rankings, draft value analysis, team grades, visualizations, and an exportable Excel report.
The objective of this project is to answer the following question:
Can pre-draft basketball production and physical measurements identify future NBA success without relying on draft position?
Rather than reproducing consensus draft order, the model evaluates players using measurable pre-draft characteristics.
This project combines multiple datasets into a single predictive modeling pipeline.
- Draft Pick
- Team
- Height
- Weight
- Age
- Points Per Game
- Rebounds Per Game
- Assists Per Game
- Steals Per Game
- Blocks Per Game
- Assist-to-Turnover Ratio
- True Shooting Percentage
- Effective Field Goal Percentage
The prediction model was trained using historical NBA Draft classes from 2012–2020.
The historical model predicts whether a player reaches a successful NBA career using pre-draft information.
For this 2026 draft value analysis, draft position was removed from the prediction features so the model could evaluate prospects based on production, efficiency, age, and physical profile rather than simply reproducing draft order.
The final model evaluates players using:
- Draft Age
- Height
- Weight
- Points Per Game
- Rebounds Per Game
- Assists Per Game
- Steals Per Game
- Blocks Per Game
- Assist-to-Turnover Ratio
- True Shooting Percentage
- Import Draft Data
- Import College Statistics
- Import Advanced Statistics
- Clean Player Names
- Merge Data Sources
- Engineer Model Features
- Predict NBA Success Probability
- Generate NBA Success Scores
- Rank Players
- Calculate Team Draft Grades
data/
2026_NBA_Draft_Analysis.xlsx
scripts/
01_prepare_2026_data.R
02_score_2026_draft.R
03_create_2026_charts.R
04_export_2026_tables.R
outputs/
data/
figures/
models/
tables/
reports/
2026_Draft_Model_Results.xlsx
The project automatically produces:
- Player Rankings
- NBA Success Scores
- Success Probability
- Biggest Draft Steals
- Biggest Draft Reaches
- Team Draft Grades
- Excel Report
- Predictive Modeling
- Logistic Regression
- Sports Analytics
- Feature Engineering
- Data Cleaning
- Data Integration
- Data Visualization
- R Programming
- Excel
- GitHub
- Random Forest comparison
- XGBoost implementation
- SHAP feature importance
- International player integration
- Summer League performance tracking
- Career outcome tracking
NBA front offices evaluate prospects using scouting, video, interviews, medical evaluations, and analytics.
This project demonstrates how historical NBA outcomes and measurable pre-draft characteristics can be combined into a predictive framework that provides an additional perspective on player evaluation.
The goal is not to replace scouting, but to support basketball decision-making through objective statistical analysis.
Danny Thompson
Basketball Analytics | Predictive Modeling | Sports Data Science
GitHub: https://github.com/DannyTData
LinkedIn: https://www.linkedin.com/in/danny-f-thompson/
MIT License