Goodreads Extravaganza: An Analysis of Trends in Reader Behavior

Posted on Aug 2, 2021

Using a dataset scraped by Austin Reese, which was itself based on a list of 50,000 books on Goodreads, I set out to better understand today's publishing and reading trends. Although this analysis is in no way comprehensive, it does provide an insightful visual exploration into today's readers. Using rating counts as a proxy for popularity and sales, we will be exploring which features maximize both rating count and average rating.

This exploration will be framed around the following 6 questions:

  1. What impact does page number have on popularity and rating?
  2. What impact does being a part of a series have?
  3. What impact do awards have?
  4. What are the most popular genres?
  5. What keywords are most popular in book descriptions?
  6. What trends are present regarding the sex of the author?

First, let's take a general glance at the data.

It appears to be normally distributed, but with suspicious spikes around the 3.5, 4, 4.5, and 5 marks. These spikes appear to be from books that weren't popular enough to get enough ratings to throw them into the 'gray area' between user-provided ratings. Let's see what happens if we remove all books with fewer than 1,000 ratings.

That's more like it. In addition to appearing much more evenly distributed, there was a substantial drop in the standard deviation (0.345 -> 0.266).  Now, let's take an overview of correlations via the heatmap.

The heatmap suggests that there may be a connection between the average rating and the number of pages the book has. It does not suggest, however, that longer book will be more popular. Let's dig a little deeper to confirm this.

What impact does page number have on book rating and popularity?

The vast majority of books seem to be centered around the 300-500 page zone. This visualization also suggests that as books get longer, the ratings generally improve. But what about popularity?

As expected, the longer books, though higher rated, are rarely popular. The majority of popular books seem to be in the 200-800 page range.

What impact does being a part of a series have?

My initial guess was that books in a series would be more popular and lower rated than their non-series counterparts. This guess turned out to be very, very wrong.

Books that are part of a series appear to be roughly equally popular and higher rated than there non-series counterparts. This was a surprising finding.

What impact do awards have?

The following is a comparision of books that have received an award (any award) and books that have not.

It seems to be that awarded books tend to be far more popular while, surprisingly, being equally rated to their non-awarded counterparts.

What genres seem to be trending?

Provided below is an admittedly crude assessment of changes in the number of book publications by genre over the past 70 years.

Though a much more granular assessment of genre trends is needed, this graph still reveals a few interesting trends; Fantasy shows an incredibly steady growth in popularity while Young Adult and Romance books have enjoyed a large growth in popularity over the past 20 years. It is also worth noting that the majority of the 'mega popular' books are in the Young Adult (Harry Potter, Hunger Games, Twilight, Divergent, etc.) and Romance (Fifty Shades of Gray) genres.

What keywords are most common in book descriptions?

This wordcloud provides an entertaining look into book descriptions over the past 50 years.

It's hard to glean anything definitive from it, but at a quick glance, it seems to be that relational words (family, mother, father, friends) and 'absolute' words (never, first, everything) are very common among book descriptions.

What trends can be found in publications by the sex of the author?

If we first look at the percent difference in publications by male vs female authors over the past 70 years, a striking trend is revealed.

From the 1950s through 2000, the majority of books published by year were primarily from male authors. Starting around 1990, this begins to shift until, in 2005, male and female authors are equally represented among publications. Now, it seems, most of the book publications are consistently from female authors for the first time in history - or, at least, in this dataset 🙂

If we again look at the percent difference in publications of male and female authors, except only including the books that have received an award, another interesting trend reveals itself.

With the exception of a notable spike in 2002, the same trend is revealed, though set back about 4 years. The turning point of equal publications of awarded books is at 2009, rather than the above 2005 turning point for equal book publications. It seems to be that the trend in awarded books by author sex is trailing about 4 years behind the same trend found in book publications by author sex.

What were our limitations?

Before I summarize the major findings of this analysis, I want to quickly note the following limitations of this study:

  • The dataset was large, but in no way comprehensive.
  • Using the number of book ratings as a proxy for sales, though useful, is hardly a sufficient replacement.
  • The genre analysis was very crude and based on a limited set of genres.
  • in addition, the genre trends were based on the number of books published in that genre by year, and was not weighted by book popularity.
  • The authors' sex was predicted based on the authors' first name.
  • This study focused on trends in the authors' sex, and did not take into account a diverse range of genders.

    Now, at last, the main findings...

  • Longer books, though higher rated, are predictably less popular.
  • Awarded books are more much popular, but similarly rated to non-awarded books
  • Series seem to be higher rated, but not much more popular than non-series books
  • Young Adult, Romance and Fantasy books seem to be trending
  • Book publications seem to be trending towards female authors
  • Awarded book publications by author sex seem to reflect the trend in book publications, though it seems to be trailing just a few years behind 

    Conclusion:

In the future, this assessment will provide, in addition to fixes to the aforementioned limitations, an analysis of the relationship between book price and book length, something this dataset did not allow.

With book publishing being such a vast, though recently challenged, industry, it is more important than ever to know what readers want. Using this and future analyses of information provided by Goodreads' vast userbase, book publishers can make wiser decisions regarding who, what, and when to publish. Now... Go read a book.

About Author

Leave a Comment

No comments found.

View Posts by Categories


Our Recent Popular Posts


View Posts by Tags

#python #trainwithnycdsa 2019 airbnb Alex Baransky alumni Alumni Interview Alumni Reviews Alumni Spotlight alumni story Alumnus API Application artist aws beautiful soup Best Bootcamp Best Data Science 2019 Best Data Science Bootcamp Best Data Science Bootcamp 2020 Best Ranked Big Data Book Launch Book-Signing bootcamp Bootcamp Alumni Bootcamp Prep Bundles California Cancer Research capstone Career Career Day citibike clustering Coding Course Demo Course Report D3.js data Data Analyst data science Data Science Academy Data Science Bootcamp Data science jobs Data Science Reviews Data Scientist Data Scientist Jobs data visualization Deep Learning Demo Day Discount dplyr employer networking feature engineering Finance Financial Data Science Flask gbm Get Hired ggplot2 googleVis Hadoop higgs boson Hiring hiring partner events Hiring Partners Industry Experts Instructor Blog Instructor Interview Job Job Placement Jobs Jon Krohn JP Morgan Chase Kaggle Kickstarter lasso regression Lead Data Scienctist Lead Data Scientist leaflet linear regression Logistic Regression machine learning Maps matplotlib Medical Research Meet the team meetup Networking neural network Neural networks New Courses nlp NYC NYC Data Science nyc data science academy NYC Open Data NYCDSA NYCDSA Alumni Online Online Bootcamp Online Training Open Data painter pandas Part-time Portfolio Development prediction Prework Programming PwC python Python Data Analysis python machine learning python scrapy python web scraping python webscraping Python Workshop R R language R Programming R Shiny r studio R Visualization R Workshop R-bloggers random forest Ranking recommendation recommendation system regression Remote remote data science bootcamp Scrapy scrapy visualization seaborn Selenium sentiment analysis Shiny Shiny Dashboard Spark Special Special Summer Sports statistics streaming Student Interview Student Showcase SVM Switchup Tableau team TensorFlow Testimonial tf-idf Top Data Science Bootcamp twitter visualization web scraping Weekend Course What to expect word cloud word2vec XGBoost yelp