The Short Version

This project is an early Tableau dashboard about streaming-platform catalogs. It combines title datasets from Amazon Prime, Disney+, and Netflix, then reshapes them into smaller analysis tables for shows, countries, genres, cast, directors, and platforms.

It is a beginner dashboard, but I like it because it already shows a good instinct: before visualization, the data needed structure.

Problem

Streaming datasets often store rich text fields like cast, countries, directors, and genres as comma-separated strings. That format is fine for reading one row, but it is weak for dashboard analysis.

The project focused on preparing the data so Tableau could answer questions like:

  • Which platform has the largest catalog?
  • How do movies and TV shows compare across platforms?
  • Which countries and genres appear most often?
  • Which actors and directors are repeated across titles?
  • How do release years and ratings vary by platform?

Approach

Data preparation - A Pandas notebook loads separate streaming datasets, assigns platform IDs, and creates a platform lookup table.

Relational reshaping - Multi-value fields are split into separate tables: casts.csv, directors.csv, genres.csv, and countries.csv. The cleaned shows.csv table keeps the title-level fields, while the supporting tables make cast, genre, director, and country analysis easier.

Dashboard build - The Tableau workbook connects to the prepared CSVs and builds a dashboard with eight worksheets around catalog composition, title type, release year, rating, country, genre, and people-related patterns.

Result

The prepared dataset includes:

  • 19,946 title records in shows.csv,
  • 114,392 cast-title rows,
  • 41,542 genre-title rows,
  • 16,334 director-title rows,
  • 12,371 country-title rows.

The result is a comparative dashboard for understanding streaming catalogs as structured data rather than isolated CSV files.

What I Learned

This project taught me that beginner dashboard work can still include serious data preparation. Splitting list-like fields into separate tables made the Tableau side much more flexible.

It also made me more aware of a common analytics pattern: good visualization often starts with small data modeling decisions.

Updated: