Book Recommendation System API

1. Introduction

This repository consists of a recommendation engine which predicts and returns the most similar books on the basis of the user's history.

2. Dataset

• For the dataset, I searched about the Project Gutenberg which is a library of over 60,000 free eBooks. It's mission is: To encourage the creation and distribution of eBooks.

• To access the free eBooks, the site has an option for offline catalogs.

• 

• In that section, there is an option 6., i.e., for The Project Gutenberg Catalog Metadata in Machine-Readable Format.

• 

• On clicking that, a link is provided for an excel-compatible CSV spreadsheet of eBook metadata. On accessing that, a pg_catalog.csv file is downloaded.

• 

• Sanitization of data-

In this, the books with incomplete data like that of author or the language are removed from the csv file.

• book_id concept-

For the ease of searching, the book elements are assigned an 'id' according to their serial numbers.

3. Build

The Recommendation System is made using Python, Flask and is based on the Cosine Similarity Algorithm.

The Prototyping is done using the Google Colab platform. Here is a working example, Book_Recommendation for the same.

4. Deployment

The server is also deployed on Heroku as Recommendation_Engine.