PMDb
Interactive web app showcasing a Snowflake database of 37,000+ films with IMDb and Letterboxd ratings
Overview
- An end-to-end medallion pipeline: raw Kaggle datasets land in Snowflake as bronze tables, dbt models clean and join them into silver and gold layers, and a Streamlit app presents the results
- Explore film rankings across multiple rating systems with leaderboards filtering by decade, popularity, and rating
- Compare IMDb vs Letterboxd lean using rating differentials to highlight where the platforms disagree most
Tech Stack
| Layer | Tools | Description |
|---|---|---|
| Ingestion | Python, Snowflake Connector | Loads raw Kaggle datasets into Snowflake as bronze tables |
| Modeling | dbt Core | Standardizes types and filters nulls into silver tables, then joins both platforms on title + year to compute composite ratings and differentials in a gold table |
| Presentation | Streamlit, Plotly | Renders interactive charts and film lookup from a periodic static export of the gold table |
Skills Developed
- Medallion (bronze/silver/gold) architecture
- Transformation modeling with dbt
- Cloud data warehousing in Snowflake