Kindle Highlights and Ratings ============================= When reading I've always underlined sentences that make me happy. Once the kids got old enough to understand there's no email or fun on a Kindle I switched from dead tree books, and now the underlining is stored in Amazon's datacenters. After a few years of highlighting on Kindle I started to wonder if the number of sentences that I liked and the eventual five-star scale rating I gave a book had any correlation. Amazon owns Goodreads_ and Kindle services sync data into Goodreads, but unfortunately highlight data isn't available through any API. I was able to put together `a little Python`_ to scrape the highlight counts per book (yay, BeautifulSoup_) and combine it with page count and rating info from the goodreads APIs. Our `family scientist`_ explained "the statistical tests to compare values of a continuous variable across levels of an ordinal variable", and there was no meaningful relationship. Still it makes a nice picture: .. attachment-image:: highlight-chart.png :width: 591px :height: 443px :alt: Highlights Per Page vs. Rating .. _a little Python: https://github.com/Ry4an/kindle-highlight-counts/blob/master/counts.py .. _family scientist: https://twitter.com/katewbauer/status/1117580683415834626 .. _Goodreads: https://www.goodreads.com/ .. _BeautifulSoup: https://www.crummy.com/software/BeautifulSoup/ .. read_more The data_ I pulled only covered books I'd made highlights in, which seems to be about 2/3rds of them. I was happy to see that more than 40% of the books I'd read and highlighted since getting a kindle were written by women, and even better than that over the last two years. That probably comes from following good people on Twitter. .. _data: https://docs.google.com/spreadsheets/d/1G2Fqs3zYlbWX5EaDTWyGiHnzvI-Jdu1ixMn3dXu0Dm4/edit?usp=sharing .. tags: ideas-built,software