A Platform to Discover, Track, and Share Your Library

WRITTEN BY: RYAN ZERNACH

SUMMARY — BetterReads is an enhanced recreation of Amazon's GoodReads. The user is able to create an account, select which genres of books they find appealing, adjust their profile, find recommendations based on a wide variety of interchangeable factors, and more! This was an extremely fun and successful group project that users can actually enjoy!

TECH STACK — AWS RDS Database, AWS Elastic Beanstalk, Docker Containerization, GoogleBooks API, Flask, FastAPI, Scikit-Learn, KNearestNeighbors, One-Hot Encoder, Natural Language Processing, JSONify, Numpy, word2vec

TEAM & TIME — Four Data Scientists, Three Web Developers, Two iOS Developers, and One UI/UX Designer, Seven Weeks

DATA SCIENCE CONTRIBUTIONS —

▻ Iteratively developed/improved Natural Language Processing (NLP) predictive models to recommend books to user

▻ Developed a Collaborative Filtering model based on user ratings (see Data Science presentation video below for more information)

▻ Facilitated the process of surveying users to determine how to improve our model — learned how to automate this process

▻ Created back-end Python API’s with Flask on AWS’s Elastic Beanstalk to facilitate JSON post requests for the web/iOS developers

▻ Launched a PostgreSQL instance on AWS RDS to store millions of GoogleBooks data, most of which was manually scraped using a multi-threading approach

▻ Collaborated on a remote team of data scientists, iOS developers, web developers, and a UI/UX designer to produce BetterReads

Data Science Presentation: 2 Months in 4 Minutes (Presented by Me)

iOS Presentation: 2 Months in 2 Minutes (Presented by Ciara)