I started my Goodreads account in 2008, when I was a college student living in Buffalo. I kept adding to the list through a move to Chicago, graduate school, my first professional job in Washington, DC, and through to my current life in an even more professional role in post-Covid New York City. I feel like a hugely different person than I was 15 years ago. Is that true? If it is, is it reflected in my reading habits as tracked through Goodreads?  

In honor of the books I haven’t read, I’ve used quilting squares from a quilt I never made to create some of my visualizations.

Data

My Goodreads list has 1,154 books on it and stretches over six eras:

• Buffalo, college student in humanities degree program (March through summer 2008);
• Chicago, non-profit gruntwork (September 2008 through August 2009),
• Chicago, graduate student in public policy program (August 2009 through summer 2011);
• Washington, DC, first professional office job as consultant (September 2011 through 2014);
• New York, second professional job with advocacy organization (2015 through February 2020); and
• New York, Covid and remote work (March 2020 through 2022).

Goodreads provides an exportable list which includes the book’s name, author, date added, date read if the book is finished, bookshelves assigned by the account owner, and three mutually exclusive bookshelves (read, to read, and currently reading). I have not assigned shelves with enough detail or consistency to be useful, so to get better categorization I uploaded the spreadsheet to Library Thing. Library Thing provides a Dewey Decimal number and detailed subject classification using the Dewey system. I added a column for location and my role at that time (student, office worker, etc.) based on date.

One challenge of the data is that I am unsure how consistently I have updated my Goodreads account over time. Another is that the lengths of time are so different for different eras. I don’t think I had a phone capable of updating my Goodreads account until at least the end of graduate school, and it is possible I didn’t have one until after I lived in DC. The longest is the time between moving to New York and the Covid lockdown (five years), and the shortest is living in Buffalo as a college student (less than a year). 

Has the ratio of books marked “want-to-read” versus finished changed over time?

When I started the Goodreads account, I had a gigantic amount of shame around not keeping up with the reading at school. My hypothesis was that I would get better at finishing books as I matured. That isn’t what really happened. A line graph showing the ratio of books I read versus books I said I wanted to read by year shows the ups and downs (you can open the full graph here):

Until 2015, the ratio is under or around one (I added more finished books than want-to-read books or nearly the same). After I moved to New York, there is an explosion of books added to my want-to-read list. I added over four times as many books to the want-to-read list during my time in New York before Covid as I actually read. I had a completely false memory of adding hundreds of books to Goodreads as escapism when I was having a hard time before moving to New York, but I really started adding them at a high rate after that, when I was in a much happier mindset. When I feel happier, I feel more curious and have more varied interests, which may mean I came across more interesting books.

I may also have lost some anxiety about not finishing books. Even though my account is not very public, I think at times I avoided adding books if I was afraid I wouldn’t really finish them. The want-to-read category is also really blunt, and includes a lot of books I’ve meaningfully engaged with. For example, Peter Ackroyd’s London is a book I’ve read sections of multiple times, and I’ve even planned travel with it. But I haven’t read every chapter, and I probably won’t. That doesn’t mean it isn’t one of my favorite books, or that it hasn’t had a big effect on how I think. After looking at the data, it’s clear that the big want-to-read spikes indicate times where I’m reading a lot and getting a lot out of it rather than moments I’m failing to follow-through in some way. Looking at this has really convinced me to stop thinking about every unfinished book as some sort of moral failure.

Is there a mismatch between the genre of books I say I want to read and the books I actually read?

I thought I would see myself adding a bunch of books for smart people who keep up with things but then actually reading nothing but Agatha Christie. That is sort of what happened. I categorized books by genre using the first digit of their Dewey Decimal numbers. I then standardized as though each era was the same length of time, and multiplied by two to try and get more space at the low end of the scale. The colors represent nine different genres, the triangles are the books added to “want-to-read”, and the squares are the number of books I actually read (you can see a larger image here): 

My reading is more varied than I though it was, and I needed a lot more markers than expected. “Literature” (the catch-all for fiction) is always first for books I’ve read, but it jostles for position with History and Geography and Social Sciences for books I want to read. For the most part, I say I want to read a lot of fiction and then I do.

This is another place that made me feel irritated at the “want-to-read” versus finished categories. There are a lot of books in the Social Science and History categories that I have obtained substantial information and inspiration from without finishing. The triangles aren’t really aspirational failures – when they are further along in the graph, I am reading a lot in that genre, it just doesn’t always make it to the finished category.

Has the subject matter of books I am interested in or read changed over time?

Word clouds could provide more nuanced information about the subject matter of the books on my list. Library Thing provides the Dewey decimal categorization for each book.  For example, John Keel’s The Mothman Prophecies is categorized as Computing and Information > Controversial Knowledge > Information > Knowledge > Mysteries > UFOs and Stephen Morris’ biography Record Play Pause is The Arts > Biography > General principles and musical forms > Music > Rock Songs and so on. I separated these into individual words and made one long list. You can look at the word clouds here. I filtered out the big genre categories like Literature, and for the sake of legibility stuck to words that appeared three or more times.

There isn’t really a lot of change in these when constructed that way. The biggest categories are always fiction, and the most prominent subjects are crime, disease, and social problems. More health-related words started to appear when I started graduate school, which was a public policy program where I did spend some time on health policy. But most of these are physician memoirs, not anything related to serious policy issues.

The word clouds do show something familiar about my taste, but they are much more interesting to me when I stop filtering for less frequent words. I started focusing on classics (still trying to make up for the reading I didn’t do in college) when the pandemic started, and I was really hoping to see evidence of that in this project so that I could tell someone about all the classics I read. But the cloud is always dominated by social problems and homicide, because a lot of those books are faster reads or things I have foundational knowledge of already. They pile up quicker than something like The Tale of Genji or The Iliad, which I spent a lot of time on (after three years, I’m almost ready to move on from antiquity to Medieval literature). There’s a hint of it in the previous visualization: in post-Covid New York, I finished more books in the Literature category than I added to want-to-read, which I think is because I was being intentional about the classics stuff. But for the most part, the most rewarding and enjoyable reading experiences I’ve had in the past three years aren’t really observable in this type of dataset. I turned the filter off in my word cloud for the present era, and that is the only way I can see those things. But that is almost the same thing as just reading the list of books.