Library of Alexandria

Inspiration

We were inspired by the social media Letterboxd, whose recommendation system for movies inspired us to do a similar project for books. We have added the chatbot implementation for more interactivity.

What it does

Alexandria chats with the user to learn more about their interests. Upon looking in the database, she recommends books that she thinks might interest the user.

How we built it

Alexandria was built using Google Cloud Natural Language. We gave Google many books to read and categorize. Then, the user's input is associated to a category, and Alexandria recommends books of this category.

Challenges we ran into

The main challenge in developing this application was to find appropriate data to categorize and to link the chatbot component to the Google Cloud system for topic detection of the user input.

Accomplishments that we're proud of

Building a simple but effective chatbot, with no previous AI knowledge, and applying advanced ML techniques to detect categories remarkably effectively.

What we learned

We learned how to use the Google Cloud services, as well as implement an API for the first time. It was also our first time dealing with artificial intelligence.

What's next for Library of Alexandria

Like any library, our plan is to always have Alexandria growing. The current database consists of 4000 books which is still quite small considering the categorization capabilities of Google's NLP API. Our first avenue for growth would be designing a web scraper that downloads and categorizes all books from library genesis. Having the capacity to recommend over 2 million books far surpasses other websites such as Goodreads since at the end of the day, they depend on user data. Alexandra is fully independent, capable of suggesting books that haven't been read in decades by anyone.