Problem

Our team was tasked with creating a movie recommendation system for a streaming service. The current recommendation system was not personalized and did not take into account the user's viewing history or preferences. This led to users receiving recommendations that were not relevant to their interests, resulting in decreased user engagement and satisfaction.

We needed to create a recommendation system that would provide personalized recommendations based on the user's viewing history and preferences while also being scalable and efficient.

Click here to see our demo.

Solution

We decided to use Cosine similarity to create our movie recommendation system. We first created a database of movies with their attributes such as genre, director, actors, etc. We then used the user's viewing history and ratings to calculate the similarity score between each movie in the database and recommended movies with the highest similarity scores.

To make the system more personalized, we added a feature where users could rate movies they had watched. This allowed us to further refine our recommendations based on their ratings and preferences.

Impact

Increased user engagement by X%
Improved customer satisfaction by X%

"This movie recommendation system has completely transformed my streaming experience. I now receive recommendations that are tailored to my interests, making it easier for me to find new movies I love."

Team

Team Member #1 - Data Scientist
- Developed the Cosine similarity algorithm and implemented it into the system
- Analyzed user data to improve recommendations
Team Member #2 - Full Stack Developer
- Built the front-end and back-end of the recommendation system
- Integrated the database of movies and user data into the system

Fake Note Detection System