Movie-revenue prediction
EDA over 45,000 movies and a regression on what actually drives revenue.
Pythonpandasscikit-learnJupyter
Exploratory analysis over Kaggle's The Movies Dataset and a regression on what actually drives box-office revenue.
Things I implemented:
- Cleaning: parsing nested genre, cast and crew JSON columns, fixing types, merging six CSVs
- EDA in pandas: genre economics, director and actor frequency, releases over time, a correlation heatmap
- Feature engineering: profit and return-percentage
- Linear regression on budget, popularity and vote count, evaluated with R squared and MAE
- Findings: popularity and vote count beat budget as predictors; a high budget does not guarantee profit; animation, family and adventure are the safer genres