Look into and explore the site: https://www.goodreads.com/ . Make a scraper to scrape the site for atleast 10 genres and atleast 5k data points for books/reviews/ratings.
List different kinds of practically useful models that you can build using this data, rate them on their practicality and also provide which use cases can we use these models the best. Then, make a recommender model of your choice using the data and optimize it.
Make a PPT explaining your approach in Part 2, with why you chose this problem statement, why you chose the model that you chose, shortcomings and future scope.