One of my favorite parts of working in a library is what we call ‘reader’s advisory.’ This is what happens when someone says something like, “I’ve read all of CJ Box’s books, what should I read next?” My first thought is that Box writes fast-paced books which take place outdoors and involve crime. I ask some questions to help me discover if the reader likes the pace, the setting or the good vs. evil parts of his books. I use that information to recommend another author for them to try.
For some reason, computers don’t have that process down yet. I recently read a book called “Galileo’s Daughter.” It was a nonfiction book about Galileo di Vincenzo Bonaiuti de’ Galilei (Galileo) and his daughter who had been a cloistered nun. The author researched the book through letters from the daughter to Galileo. Somehow 120 of her letters have survived since the 1600s. It was a good book. I added it to my Goodreads.com page, where I keep track of what I have read.
Goodreads is owned by Amazon.com. It’s programmed to recommend books based on what I have read. I generally ignore their recommendations, but this one caught me off guard. Based on my having read “Galileo’s Daughter: a historical memoir of science, faith and love,” Goodreads recommended “Daddy’s Naughty Girl.” I hate to admit I judged a book by its cover, but “Daddy’s Naughty Girl” featured a young woman…who was clearly not a cloistered nun, and I firmly hope the content was not nonfiction.
How can the computer be so wrong? My definition of “algorithm” is “the process a computer uses to guess what you would like based on data.” The Goodreads algorithm had a lot of information about what I read, nearly 2,000 books of information, yet it missed this one by a mile.
Facebook uses algorithms- Amazon and Google do too. Have you ever searched for something on Amazon and suddenly see ads for it all over your computer? In fact, every time you interact with a computer to shop or use social media it is likely using an algorithm to better understand what you might like. Algorithms choose to show you certain products by using information about what you spend time doing online.
Algorithms also help us manage information overload. They notice we don’t interact with that distant cousin on Facebook, so we see fewer of their posts. Based on our Amazon purchases they recognize that we have a dog, so they make sure we see advertisements for products our dog may like, rather than baby toys. But sometimes we stump the algorithm, and it makes an assumption based on data it doesn’t understand.
If you want to stump me, ask me to recommend a fantasy, science fiction or romance novel. Fortunately librarians have tools to help us recommend books in genres we don’t read. There are sites on the internet that don’t use Amazon’s algorithms to recommend books. A quick search on one of those sites will help me find a book that you might like.
At the Gering Library, you can get recommendations by real people who know the difference between historical nonfiction and a steamy (and maybe more than a little bit icky) romance novel. As I always say, “I didn’t write this book, so if you don’t like it, it won’t hurt my feelings.” Letting me know you didn’t care for a book will help me recommend a better book for you next time.