Reading and Rating
Every two months, I get together with the same group of guys. We’ve been meeting for over 16 years, mostly the same 15 guys. In that time, we’ve become close friends. We’ve been together through the the starting and ending of jobs, businesses and careers. We’ve shared the graduations of our kids and the births of a few grandkids. We’ve supported each other through injuries, surgeries, depression, and the deaths of family. I love these guys.
Not bad for a book club. Especially an all-guys book club.
We call ourselves The Bobos, named, tongue pressed slightly in cheek, after David Brooks’s “Bobos in Paradise.” The book title describes the lives of the Bohemian Bourgeoisie, and many years ago, after a few minutes idle contemplation over reasonably good, but affordable Shiraz or maybe a local, craft beer (probably a double IPA), we decided the description fit.
We communicate daily through an email list about topics that range deep and far beyond the books we’re reading–sports, sometimes, but less often than you’d probably think; movies, especially discussions about how we’d cast the book we just read; politics, our best and worst discussions are about politics; and gorgeous, evocative recollections or reflections of moments lived or missed, just out of reach. We talk and joke about many topics, but we are serious about the books we read. Possibly too serious.
One of things that distinguishes The Bobos from other book clubs is that we keep score. Maybe it’s a guy thing, but we’re not content merely to discuss theme, plot, character development and literary merits. We rate our picks and keep stats. We store the scores in a Google Sheet, with an offline back-up, just in case the Russians try to skew the results.
For a long time now, I’ve thought it would be fun to dig into some of this data. Recently, I started a Data Science course through UC Berkeley and edX, because I want to bolster my data science skills, create a foundation for working with AI, and earn my Junior Data Scientist badge.
I’ve been learning to use Python to manipulate data and make fairly ugly charts. I’ve also been relearning that coding is mostly about the endless search for missing commas, parentheses and quotation marks. To get anywhere close to proficient with Python for Data Science, I need more practice than the course labs provide, so I thought “why not use my nascent skills on the Bobos book ratings? Maybe I’ll uncover the secret formula for picking the perfect Bobos book.”
Start at the Top
The first thing I did was rearrange the data from the Bobos ratings sheet into a table. The table contains all of the ratings from each Bobo over the years, so it’s a lot of data. I decided to trim it down and look at just the Top 10 books and their average ratings.
Those looked right to me. And look, I’ve got one of those top-10 picks. Hmm, Busch has 2 though, and “Disgrace” with a 97.1 rating might never be topped. I decided to look at the bottom 10.
Well, crap. Guess who has THE absolute lowest-rated book on the list. At least Busch is close, pulling down his average from “Disgrace,” and Lundgren has two of the bottom 10. But “Little Known Facts” hurts. C’mon, it wasn’t that bad! Worse than “Eileen?” Nah.
Is Our Midnight’s Children Learning?
I’ve sometimes wondered if we’ve gotten any better at picking books for each other over the years. I decided to create a couple of charts to see if I could detect improvement.
Nope. Pretty much not.
The barchart on the left doesn’t seem to show any discernible pattern of improvement, and the scatterplot on the right shows only a slight improvement trend over time, but two points over sixteen years is not a lot to write home about. The Bobos are slow learners, it seems.
Now, I’ll confess that sometimes I’ve chosen books that I want The Bobos to read rather than books I thought the Bobos would like. We’re all straight, white men born within a couple of decades, living in the Midwest. A little horizon-expansion doesn’t hurt. Maybe the other guys do the same thing. Maybe they pick the books they want to read or they want us to read, instead of thinking of the books they think we’ll like. Or maybe there’s another explanation for our failure to learn from failure.
Maybe it’s because we don’t stick with hard stuff. One thing I’ve noticed from looking at our Google spreadsheet is that there seems to be a correlation between the number of Bobos who read the book and the book’s rating. The better the book, the more of us who finish it. When the book is bad, we drop it. I decided to look at the data to see if that hunch holds up.
Here’s a scatterplot of ratings and number of Bobos reading.
There seems to be a strong correlation between the number of readers and the rating of the book, but I also know that the data here is skewed a bit because our numbers were smaller in the early days. So, I ran it again starting with book #51, when we added our final member.
Seems pretty clear here, but there is an outlier with only two readers. I dug into the data, and that stinker was Salman Rushdie’s “Midnight’s Children,” which we saw earlier on the Bottom 10 list. I wondered if it was skewing the results, so I took it out it and reran the chart.
The result? The Bobos are quitters.
I was also curious to see the distribution of our picks and whether our ratings formed some sort of bell curve.
They do, sort of, with a couple of gaps. My reading from this chart is that anything below 50 is a loser that the Bobos wouldn’t recommend, and anything above 80 is a terrific read. So, with that in mind, I ran the lists again and can present an expanded Best and Worst of the Bobos:
The Bobos’ Best
The Shelf of Shame
Crap. I’ve got TWO losers? Well so does Busch. And Lundgren has three! I knew “In A Strange Room” was a mistake. The Booker Prize shortlist is a curse. Plus, Dennis always hates anything with unlikable protagonists.
What Really Matters
But, of course, all of this is just a prelude to the ultimate analysis. What we really care about knowing is: Who is winning? Who is the best Bobo book picker?
And the data says:
And there we have it, ladies and gentleman, the most excellent Bobos book picker is one of our founding members, David Earls. Earls!? How did he make it to the top of the list? Spends all his time taking pictures of birds! And Howard? That newbie?! I shouldn’t even include him. He’s only been here since book #51! At least I’m not Coons. Though, his wife does make amazing desserts, but would it kill him to bring some gluten-free once in a while?
Data Science may not help me improve my standing as a book picker, but there is more analysis and visualization I want to do. As I expand my skills, I hope to be able to find similarities and patterns in the Bobos’ rating behavior. Or possibly find a database of book attributes and see if I can create a predictor of books that the Bobos will like. Perhaps that will help me improve my score a bit, although I think “Little Known Facts” and “In a Strange Room” will likely drag my ratings down forever.