This project aims to predict the success of movies based on a range of features, including both metadata and text-based variables such as reviews. The ultimate goal was to create a model that could take into consideration various movie attributes like actors, directors, genres, and budgets, as well as sentiment from text reviews, to predict a movie's success.

The dataset comprises several features, including but not limited to:

• Metadata: Actors, Directors, Genres, Budgets
• Text Data: Reviews (processed into unigrams and bigrams)

The project employs various machine learning algorithms, including:

• Logistic Regression
• Random Forest
• XGBoost
• Ensemble Methods (Voting and Stacking Classifiers)

1. Feature Engineering: Experiment with different encoding techniques for high-cardinality features like actors and directors.
2. Neural Networks: Investigate the use of neural networks and embedding layers to capture more complex relationships among the features.

• 1-Cleaning_MovieLens_Dataset:

  • Start here. Cleaning the main MovieLens Dataset of user review text.

• 2-Cleaning_Supplemental_Metadata:

  • Cleaning supplemental movie metadata dataset.

• 3-EDA:

  • Exploratory Data Analysis and final cleaning of dataset.

• 4-Baseline_Modeling:

  • Final dataframe creation, feature engineering, text pre-processing, and baseline modeling.

• 5-Further_Baseline_Modeling:

  • Re-approaching feature engineering with different methods, re-doing text vectorization, and running further baseline models.

• 6-Final_Modeling:

  • Final modeling and hyperparameter tuning, along with model evaluation and interpretation.

• final_modeling_separate:

  • Original individual notebooks for final models, originally run in colab.

• imdb_id:

  • IMDb's unique identifier for each movie or show.

• title:

  • The title of the movie.

• actors:

  • List of main actors in the movie or show.

• directors:

  • List of directors who worked on the movie.

• genres:

  • List of genres associated with the movie.

• language_en:

  • Whether the original language was English (1) or other (0).

• year:

  • The year the movie was released.

• runtime:

  • The duration of the movie in minutes.

• budget:

  • The production budget of the movie.

• box_office_gross:

  • The total box office gross of the movie.

• production_companies:

  • List of production companies involved in the movie.

• votes:

  • The number of votes that the movie has received on IMDb.

• rating:

  • The IMDb star rating of the movie out of 10.

Note: directors, actors, production_comapnies are listed in order of "credit" (importance) as per the official movie credits. Genres are listed alphabetically.