I built a book recommender system using collaborative filtering approach. I started by cleaning and modeling the data in Jupyter Notebook. Once the model was trained, I exported it using pickle and saved it in my VS Code project folder. To provide an interactive interface, I used Flask to develop a web application. The application has two main routes: the home page and the recommendation page. On the home page, users can see the top 50 recommended books. In the recommendation page, users can enter a book name and click the 'recommend' button. The Flask route for the recommendation page receives the user's input, loads the pre-trained model from the pickle file, and generates the top 5 book recommendations based on collaborative filtering. The recommendations are then displayed on the page. I used HTML templates to render the recommendations and ensure a user-friendly interface.
The datasets I used for this project are from kaggle. Link to the datasets: https://www.kaggle.com/datasets/arashnic/book-recommendation-dataset
The landing page where the top 50 books with most amount of ratings is show.
After clicking on recommend option on header, the user will be directed to recommend page.
Then, when the user enters a book name and clicks on recommend button, top 5 similar books will be generated using collaborative filtering based approach.
