Weezer frontman Rivers Cuomo has an unusual hobby for a rock icon: fetching data from the Spotify API. He discussed his penchant for data mining in a 2022 podcast interview.
"How many bars is the average verse? I feel like my verse is a little long. What is everyone else doing? And I'll just go and analyze a hundred songs and see how many bars are in the verse. Then I realized I can keep all this stuff in Google Sheets. "I've hoarded massive amounts of data that way. It was such a pain to work out the formulas in Google Sheets that I started learning computer programming. It's a much easier way to go through all the data. I can refresh it and write Google scripts that go to Spotify and say, alright, how many streams have these thousand songs gotten? And then I can sort and see well, here are the popular songs. How many bars do the popular songs have?"
Challenged on how the results of his analysis might shift his songwriting toward popular trends and away from his artistic voice, Cuomo demurred. "To me, the fun part is writing the scripts," he said.
Turns out some celebrities really are just like us: muddling through a Python script with a coffee and a furrowed brow. Once I got over my surprise that a globally famous rock star voluntarily contends with API script bugs, authentication issues and rate limiting in his spare time, I thought about how commonly data exploration is derided in our culture as creepy, arcane or vainglorious. Certainly, data collection can be all those things and worse: exploitative, invasive or even illegal. But most of the time, in my experience, it's done by people like Cuomo, following a data owner's rules and feeding a core human need: epistemic curiosity.
Psychologists tell us we have 3 kinds of curiosity: diversive, empathic and epistemic. While all three are integral to our lives, diversive curiosity is considered the most common and least useful. We're naturally interested in anything new, a tendency which leads us to aimless scrolling, binge watching and window shopping. Empathic curiosity, the desire to understand someone's feelings, is most often healthy and helpful, but better achieved through a personal conversation than any API. It's in the care and feeding of epistemic curiosity where data mining shines as one of the better ways to spend your online time.
Epistemic curiosity is the desire to understand a topic deeply and comprehensively, to master and utilize its subject matter. Lots of experiences in the digital world foster epistemic curiosity: reference tools like Wikipedia, despite their biases; web forums featuring in-depth, expert conversations, despite their trolls; and the most informative and insightful consumer reviews all fit the bill. But mostly, the digital realm is skewed toward diversive curiosity: short-form dance videos, a friend's vacation pics, an online personality quiz. We spend time on these things not because we really want to see, but because we can't bear not to see.
I initially became interested in data mining because it saved me time in work and life. Then, I found it simply enabled me to understand a topic or community better than I could otherwise. Making better decisions with data analytics became less of a time-saver, and more of a superpower.
With superpowers comes responsibility. That's why I want to help break the stereotypes around the practice of data mining. That starts not with driving our curiosity, but with ensuring our data gathering activities are safe, sane and ethical, especially with regard to privacy, content ownership and bias recognition. It's incumbent upon us to examine these issues first, before we issue a single request for data. After that, as Cuomo knows, the fun part is writing the scripts.