🎬 Serializd TV Taste Analysis & Recommendation System

A comprehensive system to scrape your Serializd data, analyze your TV taste using AI/ML, enrich data with TMDB, and provide personalized recommendations with future tracking capabilities.

🎯 What This System Does

Current Limitations Addressed

Your original Serializd data was limited:

• ❌ No ratings or reviews (despite having 394 reviews)
• ❌ No episode-level details or watch progress
• ❌ No season counts or metadata
• ❌ Only basic show titles from 480 watched shows

What We Built

✅ Enhanced Data Scraping: Captures reviews, ratings, and detailed watch data
✅ TMDB Enrichment: Adds comprehensive metadata (genres, cast, ratings, etc.)
✅ AI Taste Analysis: Uses ML to understand your preferences and viewing patterns
✅ Personalized Recommendations: Smart recommendations based on your taste profile
✅ Future Tracking System: SQLite database to log future watches and maintain watchlists

📁 System Components

1. Data Scraping

• enhanced_reviews_scraper.py - Scrapes your 394 reviews with ratings and sentiment
• click_pagination_scraper.py - Your existing scraper for watched shows (480 shows)

2. Data Enrichment

• tmdb_enricher.py - Enriches shows with TMDB metadata (genres, cast, ratings, etc.)

3. AI Analysis

• taste_analyzer.py - Analyzes your taste using ML clustering and sentiment analysis

4. Recommendation & Tracking

• recommendation_system.py - Generates personalized recommendations and tracks future watches

🚀 Setup Instructions

Step 1: Install Dependencies

pip install pandas numpy scikit-learn matplotlib seaborn requests python-dotenv selenium webdriver-manager textblob

Step 2: Get TMDB API Key

1. Go to TMDB API
2. Create a free account and get an API key
3. Add to your .env file:

SERIALIZD_EMAIL=your_email@example.com
SERIALIZD_PASSWORD=your_password
SERIALIZD_USERNAME=morbius
TMDB_API_KEY=your_tmdb_api_key_here

Step 3: Run the Complete Pipeline

Option A: Run Everything Automatically

python run_complete_analysis.py

Option B: Run Step by Step

1. Scrape Reviews (to get your 394 reviews with ratings):

python enhanced_reviews_scraper.py

2. Enrich Data with TMDB:

python tmdb_enricher.py

3. Analyze Your Taste:

python taste_analyzer.py

4. Generate Recommendations:

python recommendation_system.py

📊 What You'll Get

1. Enhanced Data Files

• serializd_reviews.csv - Your 394 reviews with ratings and sentiment
• enriched_watched_shows.csv - 480 shows with TMDB metadata
• enriched_reviews.csv - Reviews enriched with show metadata

2. Taste Analysis

• taste_analysis.json - Comprehensive taste profile
• taste_analysis_visualization.png - Visual charts of your preferences

3. Recommendations Database

• tv_tracking.db - SQLite database with:
  • Personalized recommendations based on your taste
  • Watchlist management
  • Future watch logging
  • Statistics tracking

🎭 Taste Analysis Features

Genre Preferences

• Identifies your top genres from 480 watched shows
• Calculates genre percentages and preferences
• Uses this for future recommendations

Rating Patterns

• Analyzes your 394 reviews for rating patterns
• Converts various rating formats (numeric, letter grades, fractions)
• Identifies if you're a tough critic or generous rater

Show Characteristics

• Analyzes preference for long vs short series
• Identifies if you prefer highly-rated shows
• Analyzes network/platform preferences
• Language diversity analysis

Sentiment Analysis

• Uses TextBlob to analyze sentiment in your review texts
• Identifies positive vs negative keywords
• Determines your review writing style

ML Clustering

• Groups your shows into viewing patterns
• Uses K-means clustering on genres and metadata
• Identifies your distinct taste clusters

🎯 Recommendation System Features

Smart Scoring Algorithm

Recommendations are scored based on:

• Genre Matching (40%): How well genres match your preferences
• Rating Threshold (30%): Preference for highly-rated shows
• Popularity (20%): Balance of popular vs niche content
• Recency (10%): Preference for newer shows

Data Sources

• TMDB Recommendations: Based on shows you've watched
• Trending Shows: Current popular content
• Genre-based: Shows matching your preferred genres

Tracking Features

• Watch Logging: Log episodes/seasons with ratings and reviews
• Watchlist Management: Prioritized list of shows to watch
• Statistics: Track your viewing habits over time

📈 Usage Examples

View Your Taste Analysis

from taste_analyzer import TVTasteAnalyzer

analyzer = TVTasteAnalyzer()
analyzer.load_data()
profile = analyzer.generate_taste_profile()
analyzer.print_summary()

Get Personalized Recommendations

from recommendation_system import TVRecommendationSystem

system = TVRecommendationSystem()
system.generate_recommendations()
system.print_recommendations(limit=10)

Log a Watch

system.log_watch("Breaking Bad", season=1, episode=1, rating=9, review_text="Amazing pilot episode!")

Manage Watchlist

system.add_to_watchlist("The Bear", priority=8, notes="Heard great things about this")
system.print_watchlist()

🔍 Sample Analysis Output

🎬 YOUR TV TASTE ANALYSIS SUMMARY
============================================================
📊 Total Shows Watched: 480
📝 Total Reviews Written: 394

🎭 TOP GENRES:
  1. Drama: 156 shows (32.5%)
  2. Comedy: 98 shows (20.4%)
  3. Crime: 67 shows (14.0%)
  4. Thriller: 45 shows (9.4%)
  5. Sci-Fi: 34 shows (7.1%)

⭐ RATING PATTERNS:
  Average Rating: 7.8/10
  Total Rated Shows: 394

💡 KEY INSIGHTS:
  1. Your top 3 favorite genres are: Drama, Comedy, Crime
  2. Drama makes up 32.5% of your watched shows
  3. You tend to rate shows highly, suggesting you're selective about what you watch
  4. You prefer longer series with multiple seasons
  5. You tend to watch critically acclaimed shows
  6. Your reviews are generally positive and enthusiastic
  7. Your largest viewing pattern centers around Drama shows

🎯 Addressing Original Limitations

✅ Ratings & Reviews

• Before: No ratings despite 394 reviews
• After: Full sentiment analysis and rating extraction from all reviews

✅ Episode-Level Details

• Before: No episode counts or watch progress
• After: TMDB provides episode counts, seasons, and runtime data

✅ Rich Metadata

• Before: Only show titles
• After: Genres, cast, crew, networks, languages, popularity, ratings, keywords

✅ Historical Context

• Before: Limited to Jan 1, 2022+ data
• After: TMDB provides full show history and context

✅ Future Tracking

• Before: No way to track future watches
• After: Complete SQLite system for ongoing tracking

🔮 Future Enhancements

Potential Additions

1. Integration with Other Platforms: Import from Trakt, IMDb, etc.
2. Social Features: Compare taste with friends
3. Advanced ML: Deep learning for better recommendations
4. Web Interface: Flask/Django web app
5. Mobile App: React Native or Flutter app
6. Export Features: Generate reports, share taste profiles

API Integrations

• Trakt.tv: For historical data pre-2022
• IMDb: For additional ratings and reviews
• JustWatch: For streaming availability
• Rotten Tomatoes: For critic vs audience scores

🛠️ Technical Architecture

Data Flow

Serializd Scraping → TMDB Enrichment → AI Analysis → Recommendations → Future Tracking

Technologies Used

• Scraping: Selenium WebDriver
• Data Processing: Pandas, NumPy
• Machine Learning: Scikit-learn (K-means, TF-IDF, Cosine Similarity)
• Sentiment Analysis: TextBlob
• Visualization: Matplotlib, Seaborn
• Database: SQLite
• API: TMDB REST API

File Structure

serializd-py/
├── enhanced_reviews_scraper.py    # Scrape reviews & ratings
├── tmdb_enricher.py              # Enrich with TMDB data
├── taste_analyzer.py             # AI taste analysis
├── recommendation_system.py      # Recommendations & tracking
├── click_pagination_scraper.py   # Original watched shows scraper
├── final_watched_shows.csv       # Your 480 watched shows
├── serializd_reviews.csv         # Your 394 reviews (generated)
├── enriched_watched_shows.csv    # Shows + TMDB data (generated)
├── taste_analysis.json           # Your taste profile (generated)
├── tv_tracking.db                # Future tracking database (generated)
└── .env                          # Your credentials

🎉 Getting Started

1. Clone/Download all the Python files
2. Set up your .env file with credentials and TMDB API key
3. Install dependencies: pip install -r requirements.txt
4. Run the enhanced reviews scraper: python enhanced_reviews_scraper.py
5. Enrich with TMDB: python tmdb_enricher.py
6. Analyze your taste: python taste_analyzer.py
7. Generate recommendations: python recommendation_system.py

Your goal of analyzing your TV taste and building a personalized recommendation system is now fully achievable with this comprehensive solution! 🎬✨