Movie & TV Metadata Master Database

Title

Build Comprehensive Global Movie & TV Metadata Database for Recommendation Engine

Project Overview

I am building a personal recommendation engine that predicts my rating for movies and TV shows based on a large history of titles I have already rated.

To support accurate predictions, I need a global media metadata database that contains rich structured information for movies and TV series.

The dataset should combine multiple trusted sources and be designed for machine-learning comparison against my rating history.

Scope of Work

Build a master dataset containing global movie and TV metadata.

This will serve as the candidate pool for prediction models.

The database should include titles from:
• IMDb official dataset
• The Movie Database API
• Optional enrichment from JustWatch

Required Data Fields

Each title should include as many of the following as possible.

Core identification
• imdb_id
• tmdb_id
• title
• original_title
• year
• type (movie / series / episode / miniseries)

Basic metadata
• genres
• runtime
• language
• country
• release date

Creative team
• director(s)
• writer(s)
• top cast (first 10 billed)

Ratings and popularity
• IMDb rating
• IMDb vote count
• TMDb rating
• TMDb vote count

Narrative metadata
• plot summary
• keywords
• themes/tags if available

Structural attributes
• franchise/series linkage
• episode relationships for TV
• sequel/prequel relationships

Production information
• production companies
• budget
• revenue (if available)
• streaming availability (JustWatch)

Deliverables
1. Merged master dataset
• CSV or PostgreSQL database
2. Schema documentation
3. Data cleaning
• remove duplicates
• normalize titles
• unify IDs
4. ETL pipeline
• scripts to refresh the dataset monthly
5. Matching keys
• imdb_id
• tmdb_id

Database Size Expectations
Movies: 600k +
TV Series 200k +

Technical Requirements

Preferred stack:
• Python
• Pandas
• PostgreSQL or SQLite
• API integration
• ETL scripting

Important Constraints

Do NOT scrape IMDb pages.

Use official datasets and APIs only.

Goal of the Project

The database will be used to:
• compare against a personal movie rating history
• calculate similarity between titles
• generate predicted ratings
• identify highly compatible unseen movies

Accuracy of metadata is critical.

Ideal Candidate

Experience with:
• media datasets
• ETL pipelines
• Python data engineering
• IMDb or TMDb APIs
• building large datasets