Optimization of Wheel of Fortune Bonus Round

Axyl Fredrick | Math 37700 | Spring 2025

Project Overview

Problem Area

In the final bonus round of the game show Wheel of Fortune, a contestant has to pick a category and solve a Hangman-esque puzzle. The letters R, S, T, L, N, and E are all chosen already, and the contestant has to pick between three additional consonants and one vowel, and then solve the puzzle. For this project, we will look for the best letter combination in general case scenarios and across different categories. We also intend to build a regression model that provides the best combination to maximize the letters in a given puzzle also based on the preexisting placement of the letters R, S, T, L, N, and E.

Key takeaways include:

• Prediction of what letters are most frequent
• Prediction of what letters are best in certain contexts
• Maximizing the number of letters that are covered on the board.

Affected People

• Those who decide to enter Wheel.
• People at home or school playing Wheel-like games.
• The production companies behind Wheel, as they can tailor the show to make easier or more difficult puzzles based on these metrics.

Impact

• By finding the optimal combination of letters by category and in a general case scenario, a skilled contestant who applies these findings is closer to winning tens of thousands of dollars per contestant.

Proposed Solution

• We intend to build a regression model using SciKit Learn to find which letters should be guessed in different puzzles, and find a general case guess based on what works for each word. We will evaluate whether linear or logistic works better.

• We proxied earlier based on frequencies that G, H, D, A are the best combination in terms of frequency, but we want to take into account the individual words that make up each puzzle.

Data Dictionary

• For the Wheel dataset:

Variable	Type	Description
Puzzle	String	An individual puzzle from an episode.
Category	String	The category to which a puzzle belongs.
Date	String/Date Object if possible to convert	The date that an episode aired.

• For the Word Frequency dataset:

Variable	Type	Description
Word	String	A word that is found in an online corpus of English text.
Count	Integer	The frequency of a word.

Organization

Notebooks

Path	Purpose
notebooks/1_data_prep.ipynb	Data ingestion, cleaning, and saving processed files
notebooks/2_eda.ipynb	Visual and statistical exploration of cleaned data
notebooks/3_pre-processing.ipynb	Preprocessing and testing with basic models
notebooks/4_modeling.ipynb	Advanced modeling, optimization, and final evaluation

---

Installation & Setup (OPTIONAL)

1. Clone the repo
   git clone https://github.com/YourUsername/CapstoneProject.git
   cd CapstoneProject

2. Create environment & install dependencies
   python3 -m venv .venv
   source .venv/bin/activate # macOS / Linux
   .venv\Scripts\activate # Windows
   pip install -r requirements.txt

3. Directory permissions
   Ensure data/ and outputs/ are writable.

---

Project Structure

CapstoneProject/
├── README.md
├── .gitignore
├── requirements.txt
│
├── data/
│   ├── raw/                # Immutable source data
│   └── processed/          # Cleaned, feature‑engineered data
│
├── notebooks/              # Jupyter notebooks
│   ├── 1_data_prep.ipynb
│   ├── 2_eda.ipynb
│   ├── 3_baseline_models.ipynb
│   └── 4_final_modeling.ipynb
│
├── src/ (OPTIONAL)         #  Reusable Python modules
│   ├── __init__.py
│   ├── preprocessing.py
│   ├── features.py
│   ├── modeling.py
│   └── evaluation.py
│
├── slides/                 # Slide deck PDF & source
│   ├── Capstone_Final_Presentation.pdf
│
├── outputs/                # Model artifacts & figures
│   ├── models/
│   └── figures/
│
└── tests/                  # (Optional) unit tests for src/

---

Project Flow

Raw Data Ingestion  →  Data Cleaning  
Data Cleaning       →  Feature Engineering  
Feature Engineering →  Exploratory Data Analysis  
EDA                 →  Baseline Modeling  
Baseline Modeling   →  Advanced Modeling & Optimization  
Advanced Modeling   →  Model Evaluation & Interpretation
Evaluation          →  Deployment & Reporting  

1. Raw Data Ingestion – load source files and validate schema.
2. Data Cleaning – handle missing values, remove duplicates, standardize formats.
3. Feature Engineering – create new variables, encode categoricals, scale numerics.
4. Exploratory Data Analysis – generate summary statistics and key visualizations (e.g. histograms, boxplots, scatterplots).
5. Baseline Modeling – simple models (e.g., linear regression, logistic regression, decision tree, ARIMA) to set performance benchmarks.
6. Advanced Modeling & Optimization – hyperparameter tuning, ensembles, or complex architectures (e.g. Random Forests, SARIMAX, Simple Neural Networks, LSTM, CNN, XGBoost etc.).
7. Model Evaluation & Interpretation – compare metrics (e.g. MSE, MAE, Accuracy), confusion matrix, plot ROC / precision‑recall.
8. Deployment & Reporting

---

Repository

• data

  • contains link to copy of the dataset (stored in a publicly accessible cloud storage)
  • saved copy of aggregated / processed data as long as those are not too large (> 10 MB)

• model

  • joblib dump of final model(s)

• notebooks

  • contains all final notebooks involved in the project

• docs

  • contains final report, presentations which summarize the project

• references

  • contains papers / tutorials used in the project

• src

  • Contains the project source code (refactored from the notebooks)

• .gitignore

  • Part of Git, includes files and folders to be ignored by Git version control

• conda.yml

  • Conda environment specification

• scrape_wheeldata.py

  • Program to scrape necessary episode data from HTML pages of Wheel of Fortune Compendium

• README.md

  • Project landing page (this page)

• LICENSE

  • Project license

---

License

This project is licensed under the MIT License – see LICENSE for details.

Dataset

• Source: Wheel of Fortune Compendium
• Raw File: Located under data/raw/

File name	Description
data/obtained/wheeldata.csv	full dataset scraped

Credits & References

https://www.kaggle.com/code/chrisloop89/english-words-and-letters-frequency

https://saturncloud.io/blog/how-to-iterate-through-specific-columns-and-rows-in-pandas-dataframe-to-perform-a-check/

Dataset Source: https://buyavowel.boards.net/page/compendium