Book recommender system based on staff recommendations from Powell's Books.
Powell's Books is an independent bookstore chain, based in Portland, Oregon. It was founded in 1971 and has 4 locations in the Portland Metro area. The company also operates a popular website at Powells.com.
A mainstay of Powell's are the staff picks. Picks are ubiquitous throughout the stores and popular with shoppers browsing for books. The card include a few handwritten sentences, often with drawings. Picks have also been added to Powell's website and organized by staff member.
This project includes several components to produce a recommender system.
- A SQLite database of staff recommendations, book/author details, reviews, and other info
- Scripts using Beautiful Soup 4 to scrape data from Powell's website
- Script to upload dataset to Kaggle.com
- Jupyter notebooks with EDA and machine learning code
- Implementation of the recommender model using Streamlit
Wikimedia Commons: Cacophony / CC BY 3.0
Powell's Books flagship bookstore at 1005 W Burnside St., Portland, OR 97209
Wikimedia Commons: Another Believer / CC BY-SA 3.0
An example of Powell's staff picks
Powell's website has a section of book recommendations organized by the staff members' first name and last initial. Each pick has a short blurb of why they are recommending the book.
At the end of each year, staff create members create a top five ranking of their recommended books. The top book has a short blurb on why the book is the top pick of the year.
This software is released under the MIT License. https://opensource.org/licenses/MIT