The Short Version

A focused, fast-paced EDA of the Netflix Movies & TV Shows dataset answering seven clean questions about the catalog’s composition. Notable for an early appearance of the RepetitionDF() function — the same multi-value list aggregation utility that became a recurring tool in the Google Play Store and IMDB projects — used here to count cast member and director appearances across comma-separated fields.

Problem

What does Netflix’s catalog actually look like? How are movies and TV shows distributed across release years, months, countries, ratings, cast, directors, and genres?

Approach

Seven questions, one chart each:

1. Movies vs. TV Shows ratio — Pie chart; movies dominate the catalog.
2. Release year trends — Grouped bar chart (post-1970) by content type; Netflix’s original production ramp-up is clearly visible in recent years.
3. Monthly additions — Grouped bar chart by month; content additions spike in certain months, with July and December standing out.
4. Content ratings distribution — Pie chart; TV-MA and TV-14 account for the large majority of content.
5. Top production countries — Country column split on commas (first country retained); countries with fewer than 100 titles filtered out. Grouped bar chart reveals the USA’s overwhelming dominance, with India and UK as distant second and third.
6. Most-appearing cast members — cast column split on commas → RepetitionDF() → top 20 bar chart.
7. Most-appearing directors — Same pipeline on director column.
8. Most common genres — Same pipeline on listed_in (multi-genre comma-separated) → top 20 bar chart.

Result

• Movies make up ~70% of the catalog; TV shows ~30%.
• USA produces by far the most Netflix content; India and UK are the only other countries clearing 100+ titles in the filtered set.
• TV-MA is the most common content rating — Netflix’s catalog skews toward adult audiences.
• The top cast members by appearances are predominantly actors with long-running TV shows or recurring Netflix originals.
• International Movies and Dramas are the two most common genre tags.

What I Learned

This is one of the earlier EDA notebooks in the portfolio, and it shows the RepetitionDF() function appearing in its initial form — the same utility that was later extended in the IMDB deep analysis and Google Play Store projects to handle weighted aggregation, percentage sunburst charts, and multi-key lookups. Starting simple and reusing the function across progressively more complex projects is a more effective pattern than rebuilding it from scratch each time. The country filtering step (dropping countries with fewer than 100 titles before plotting) is also a small but important data cleaning decision — without it, dozens of single-entry countries crowd the chart and make the dominant patterns unreadable.

Links

• Kaggle