ongaunjie1 / movie-recommendation-app

Simple Movie recommendation app

Movie Recommendation app using cosine similarity (Content-based filtering)

• Dataset used: https://www.kaggle.com/datasets/ahsanaseer/top-rated-tmdb-movies-10k
• themoviedb API: https://developer.themoviedb.org/reference/intro/getting-started
• Movie poster extracted from the API

General steps:

1) Data Preprocessing:

• Ensure that your dataset is clean and handle any missing values if necessary.

2) Feature Engineering:

• Use CountVectorizer on the "genre" and "overview" columns to convert them into numerical representations.

3) Creating a Feature Matrix:

• Concatenate the numerical representations obtained from CountVectorizer with other relevant numerical features to create a feature matrix.

4) Cosine Similarity:

• Calculate the cosine similarity between movies based on their feature matrix. This can be done using the cosine similarity function from scikit-learn or other libraries.

5) Recommendation Generation:

• Identify top 5 movies with the highest cosine similarity to the user's movie selection as top recommendations.

Streamlit-app: