Creating Movie Magic...
IMDb Movie Rating Prediction
Introduction
This project leverages statistical modeling to predict IMDb ratings for upcoming blockbuster movies. By analyzing data from over 2,000 movies, the model identifies key factors influencing audience ratings and provides insights into future audience preferences.
Key Highlights
- 🎯 Purpose: To predict IMDb ratings for 12 major blockbusters releasing in November 2023.
- 🔍 Technology: R for statistical modeling, ggplot2 for visualization, and advanced regression techniques for predictive accuracy.
- 📊 Results: Achieved low Mean Squared Error (MSE) for IMDb predictions, with robust cross-validation metrics.
- 💡 Challenges: Tackled issues like heteroskedasticity, multicollinearity, and skewed predictors while optimizing model complexity.
Dataset
The dataset consists of key features from over 2,000 movies, including:
- IMDb ratings as the target variable.
- Film characteristics: budget, duration, release date, etc.
- Cast details: lead actors and IMDb star meter rankings.
- Production details: director, distributor, cinematographer, etc.
- Dummy variables for genres and maturity ratings.
Modeling Approach
Before training the model, data was preprocessed to ensure consistency:
- Data Cleaning: Removed invalid entries and missing values.
- Feature Engineering: Dummified categorical variables (genres, ratings).
- Normalization: Applied Min-Max scaling for numeric predictors.
- Outlier Handling: Detected and handled extreme values using Z-scores.
A robust linear regression model formed the backbone of this project, enhanced with:
- Polynomial Regression: Captured non-linear relationships in predictors.
- Splines: Improved fit for variables like budget and star rankings.
- Interaction Terms: Explored dependencies between genres and budget.
- Validation: Performed k-fold cross-validation to ensure model robustness.
The following metrics were used to evaluate the model:
- R-Squared: Assessed goodness-of-fit for the training data.
- MSE: Measured prediction error on validation datasets.
- Residual Analysis: Checked for non-linearities and heteroskedasticity.
Visualizations
Visualizations include bar plots, scatter plots, and line charts explored during Exploratory Data Analysis.
Predictions
Using the finalized model, IMDb ratings for 12 upcoming movies in November 2023 were predicted. Hover over the posters to reveal the IMDb scores!
Pencils vs Pixels
IMDb Rating: 6.8
The Dirty South
IMDb Rating: 6.1
The Marvels
IMDb Rating: 6.5
The Holdovers
IMDb Rating: 6.5
Next Goal Wins
IMDb Rating: 7.5
Thanksgiving
IMDb Rating: 7.1
The Hunger Games: The Ballad of Songbirds and Snakes
IMDb Rating: 5.4
Trolls Band Together
IMDb Rating: 6.4
Leo
IMDb Rating: 6.9
Dream Scenario
IMDb Rating: 6.8
Wish
IMDb Rating: 7.0
Napoleon
IMDb Rating: 8.8
Key Insights
Genre Influence
Documentaries (+1.023) and Animations (+0.937) excel, while Horror (-0.538) and Action (-0.309) underperform. This reflects audience preferences and genre-specific appeal.
Seasonal Advantage
Movies released in November scored +0.6 higher than those in summer and spring, driven by family audiences.
Budget Insights
Films with budgets exceeding $50M tend to score an average of 0.8 points higher, showing a strong correlation between investment and quality perception.
Star Power
Lead actors with IMDb Star Meter ranks less than 100 significantly boost ratings, adding an average of +0.7 to a movie's score.
Cinematography
Expert cinematographers with more than 6 movies boost average ratings by +0.5, highlighting the importance of visual storytelling.
Runtime Insights
Movies with runtimes between 90-120 minutes scored better than overly long or short films, emphasizing the importance of optimal pacing.
Additional Resources
Explore the full project, detailed codebase, and additional insights on GitHub!
Visit GitHub Repository