I love to read books, especially mystery books by my favorite authors (Louise Penny, Fred Vargas, Kate Atkinson, Nancy Atherton and Jussi Adler-Olsen). They are all authors who have crafted a wonderful cast of thoughtful characters in picturesque settings. I live in Seattle, which is also picturesque (especially in the summer) but I do enjoy considering what life must be like in Quebec, France, England and Denmark.
Since this group may only reasonably create and publish 1 book a year, I do have a larger group of authors in my reading portfolio because I may read a new book every 2-3 weeks if book releases go my way. Unfortunately I can’t always depend on a book to be my cup of tea because a story may be situated in an unpleasant setting (hunting lodge) or could be a new stand-alone novel that lacks any of my familiar heroines and heros.
To help me stay up to date on my favorite authors, I have developed a shiny app, myBookQ that queries the google books api to get the newest books by author. Included with author parameters are positive and negative words as well as words that may indicate a book is part of a series by that author. I am not a fan of true crime or psychological dramas. I also would prefer that innocent cats, dogs, and other creatures, while fictional, not be harmed in a story as you will see by the default word values. The analysis is currently based on the descriptions of the book provided by publishers. I would like to include in the future, reviews by professionals and other readers.
To myBookQ!
Without further ado, here is the first iteration where I return new books with the default best mystery authors ever. For the sentiment scores I am using the afinn sentiments dataset, while the comparison cloud features the bing sentiments dataset. The anticipation index is made up of the inputs where both positive and continuing series words are assigned a score of 1 and negative words, a value of -1. I will continue to refine my scoring as I apply it to books that I have read recently and regretted as well as ones that were delightful. I think it’s interesting to include wordclouds to quickly see what a book may be about and hopefully see my favorite characters present. Using sentiment analysis for mysteries doesn’t make a lot of sense due to the nature of murder, mystery and mayhem, but since this application can be used for any author, it may help vet a future purchase of books that I would prefer to be more positive than negative.
Resources & Appreciations
- Seattle Mystery Books - the original champions of my favorite authors
- Text Mining with R: A Tidy Approach by Julia Silge and David Robinson
- Google Books API
- myBookQ repo