Adjusted Metacritic Scores

A webapp that displays Metacritic movie scores adjusted for critic bias and sample size.

Live Demo: your-app.vercel.app (update after deployment)

Features

• In Theaters: Movies from the last 60 days, ranked by adjusted score
• All Movies: Browse 20,000+ movies with search, genre filter, and year range
• Bias Adjustment: Accounts for individual critic tendencies (some rate higher/lower than average)
• Bayesian Shrinkage: Films with few reviews are pulled toward the overall average
• Daily Updates: Automatic data refresh via GitHub Actions

Methodology

Raw Metacritic scores can be biased by which critics review a film. This project:

1. Estimates critic effects: Using exponentially-weighted averages, we calculate each critic's typical deviation from the metascore
2. Applies hierarchical shrinkage: Critics with few reviews are shrunk toward their outlet's average
3. Adjusts review scores: Each review is adjusted by subtracting the critic's estimated bias
4. Re-averages: Adjusted scores are computed from bias-corrected reviews
5. Applies Bayesian shrinkage: Films with few reviews are pulled toward the time-varying grand mean

This results in scores that better reflect the quality signal in reviews rather than who happened to review the film.

Tech Stack

• Frontend: React + Vite + Tailwind CSS
• Data Pipeline: Python (pandas, numpy, scipy)
• Hosting: Vercel (free tier)
• Automation: GitHub Actions

---

Deployment Guide

Prerequisites

• GitHub account
• Vercel account (free at vercel.com)
• Node.js 18+ installed locally
• Python 3.9+ installed locally

Step 1: Set Up the Repository

1. Create a new GitHub repository (e.g., metacritic-adjusted)

2. Clone it locally:

   git clone https://github.com/YOUR_USERNAME/metacritic-adjusted.git
   cd metacritic-adjusted

3. Copy the webapp files into your repo (all files from this project)

4. Create the data directory structure:

   mkdir -p data/metacritic_data
   mkdir -p data/hierarchical_effects  
   mkdir -p data/adjusted_scores
   mkdir -p public/data

5. Copy your existing data files:

   cp /path/to/metacritic_data/*.csv data/metacritic_data/
   cp /path/to/hierarchical_effects/*.csv data/hierarchical_effects/
   cp /path/to/adjusted_scores_v2/*.csv data/adjusted_scores/
   cp /path/to/webapp_data/*.json public/data/

Step 2: Test Locally

1. Install dependencies:

   npm install

2. Start the dev server:

   npm run dev

3. Open http://localhost:5173 in your browser

Step 3: Deploy to Vercel

1. Go to vercel.com and sign in with GitHub

2. Click "Add New Project"

3. Import your metacritic-adjusted repository

4. Configure the project:

   • Framework Preset: Vite
   • Build Command: npm run build
   • Output Directory: dist
   • Install Command: npm install

5. Click "Deploy"

6. Wait for the build to complete (usually 1-2 minutes)

7. Your app is now live at https://your-project-name.vercel.app!

Step 4: Set Up Daily Updates

The GitHub Actions workflow (.github/workflows/daily-update.yml) will:

• Run at 6 AM UTC every day
• Re-compute adjusted scores from your data
• Commit the updated JSON files
• Trigger a Vercel re-deployment

To enable:

1. Make sure your Python scripts are in the scripts/ directory:

   cp /path/to/compute_hierarchical_effects.py scripts/
   cp /path/to/metacritic_adjusted_scores_v2.py scripts/
   cp /path/to/csv_to_json.py scripts/
   cp /path/to/fix_dates.py scripts/

2. Push to GitHub:

   git add .
   git commit -m "Initial commit"
   git push origin main

3. The workflow will run automatically on schedule, or you can trigger it manually:

   • Go to your repo on GitHub
   • Click "Actions" tab
   • Select "Daily Data Update"
   • Click "Run workflow"

Step 5: Custom Domain (Optional)

1. In Vercel, go to your project settings
2. Click "Domains"
3. Add your custom domain
4. Follow the DNS configuration instructions

---

Project Structure

metacritic-adjusted/
├── .github/
│   └── workflows/
│       └── daily-update.yml    # GitHub Actions workflow
├── data/
│   ├── metacritic_data/        # Raw movies.csv and reviews.csv
│   ├── hierarchical_effects/   # Computed critic/outlet effects
│   └── adjusted_scores/        # Final adjusted_scores.csv
├── public/
│   └── data/                   # JSON files served by the webapp
│       ├── movies_all.json
│       ├── movies_recent.json
│       └── metadata.json
├── scripts/
│   ├── daily_update.py         # Main update orchestration
│   ├── compute_hierarchical_effects.py
│   ├── metacritic_adjusted_scores_v2.py
│   ├── csv_to_json.py
│   └── fix_dates.py
├── src/
│   ├── App.jsx                 # Main React component
│   ├── main.jsx                # Entry point
│   └── index.css               # Tailwind styles
├── index.html
├── package.json
├── vite.config.js
├── tailwind.config.js
├── postcss.config.js
└── README.md

---

Updating Your Data

Manual Update

# 1. Update your raw data (movies.csv, reviews.csv)
# 2. Run the pipeline
python scripts/daily_update.py \
  --data-dir ./data/metacritic_data \
  --effects-dir ./data/hierarchical_effects \
  --scores-dir ./data/adjusted_scores \
  --webapp-dir ./public/data \
  --scripts-dir ./scripts

# 3. Test locally
npm run dev

# 4. Commit and push
git add .
git commit -m "Update data"
git push

Adding New Movies

To add newly scraped movies:

1. Append new rows to data/metacritic_data/movies.csv
2. Append new reviews to data/metacritic_data/reviews.csv
3. Run the update pipeline (effects will be recomputed)

---

Troubleshooting

Build fails on Vercel:

• Check that all dependencies are in package.json
• Ensure public/data/ contains the JSON files

Data not loading:

• Check browser console for errors
• Verify JSON files are valid: python -m json.tool public/data/movies_all.json

GitHub Actions failing:

• Check the Actions tab for error logs
• Ensure Python scripts have no syntax errors
• Verify data file paths are correct

---

License

MIT License - feel free to use this for your own projects!

Credits

• Data sourced from Metacritic
• Methodology inspired by Bayesian hierarchical models and James-Stein estimation