The task: get a catalogue of 5,850 streaming titles (3,744 movies and 2,106 shows) ready for analysis, and understand what's in it.
What I did: profiled every column with pandas, handled missing values by what they mean, and exported a clean file for further work.
What the profiling found
- Missing values carry meaning. All 3,744 movies have no "seasons" value, which is correct rather than missing data, so it shouldn't be filled.
- Age certification is missing for 45% of titles (2,619), too many to impute, so it's kept as "unknown".
- Ratings have gaps: 482 titles lack an IMDb score and 311 lack a TMDB score.
- The catalogue spans 63 release years and 452 combinations of production country.
Skills
- Profiling with info, describe, nunique and isnull
- Deciding when to fill, flag or leave missing values
- Duplicate checks and exporting a cleaned dataset