"How did you get into data science?" I ask at the start of my interview with Beverly Shih, a data scientist at Songtradr. "We can start by talking about how I didn't get into data analytics for music," she says, laughing. "It's a typical path for people: they start [as] a data scientist, taking their hard skills, and learn about the industry after. I look at my own journey and realize I didn't do that at all."
Having always been interested in music, Beverly began her journey in music performance. As she recalls, she wasn't very good at it, which made her think she couldn't build a career in the music industry. When it came time to choose a major, Shih decided to go with biology at UCLA instead.
On top of having a robust biology degree, UCLA also has a renowned music industry program. "I had no idea what it was, so I took a class and realized that's where I want to be," Beverly recounts, admitting that she still couldn't figure out what she could contribute to the industry but knew that the industry could contribute to her.
As part of the requirement for the minor, Shih had to intern at a music company. Sending a number of applications to top businesses in the field, she ended up landing an internship at ASCAP in data analytics. "That checked a lot of boxes for me – I could engage in music without having to actually be good at performing," she says. She was able to use the hard skills she learned in biology and apply them to the industry she loved. As it turned out, it was her major that made her stand out from the pool of applicants: while everyone was applying from a music business standpoint, ASCAP wanted someone with a STEM-heavy background.
There, Beverly was busy with open-ended research projects – she was given a dataset and had to figure out what was going on there. Usually, these datasets would come from sources like Nielsen Music or client statements, and the data scientist had to parse, clean the data out, and build visualization charts. "That was the most exciting part because it was the essence of data science: taking a bunch of numbers and turning them into concepts and ideas on how people can run their business better or understand trends of the past," she explains.
Looking back, Shih understands that this was a pivotal internship. She got to wet her feet in what it's like working in data analytics for a music giant and the internship also opened her eyes to the fact that while she had a general understanding of statistics, she had to master different business intelligence tools and learn more about queries and unstructured data. "It was also about understanding how I fit into ASCAP and the music industry," the data scientist adds.
At ASCAP, Beverly was working right next to the film music department, whom she talked to often, asking questions and helping out with a few projects. Having always loved film music, she was able to look at cue sheets, composer statements, and get more accustomed to the scene. Enjoying the process, she decided to apply for film music internships next, landing one at NBCUniversal in music publishing. "It was there that, surrounded by creative instead of STEM-heavy people, I felt like I wasn't contributing as much as when I was at ASCAP doing analytics," Beverly conveys. "That's when I [knew] that I'm definitely not going to be a film composer and I could be providing more quantitative value in a quantitative role."
When the data scientist began applying for a full-time job after graduation, she stuck to more STEM-heavy roles. Seeing an open position for Music Operations Coordinator at Songtradr, Beverly didn't quite know what that meant but chose to send her resume anyway. "It turned out that a lot of my hard and soft skills aligned with that position," she says.
Although she was more in the music operations scene when she first got hired, she was doing a lot of coordination and thus was later promoted to specialist and senior specialist. "I had the data analytics lead bit tagged to my title but it never felt like I was making a major step up – it was more that I was already doing these things and my job title [was] catching up and evolving with me," Shih elaborates. Over time, there was an increase in the responsibilities and power to take initiative on projects.
"I was taking more leadership and projects because I had that technical background that some of my co-workers didn't and I was able to share that with them. I became the data analytics point person in our team."
With a shift in her department's structure, Beverly shifted to royalty processing, which was completely different from what she was doing before. While physically the job looked the same – in her words, she just stands on her feet behind a desk for eight hours – now she has to process much more data. "[I'm working] on a project that was the motivation behind my longest R script, [which] processes a huge amount of money quickly," she says.
Essentially, there are recordings, the royalties of which need to be made sense of and passed on to clients. "In this case, we were dealing with YouTube royalties," Shih explains. Originally stuck with the problem of not being able to even look at the data to do anything with it, Beverly took on the project as she figured out a way to process that data with R.
Talking about her journey so far, Beverly says she's very early in the process of figuring out how to apply analytics in the music industry, adding that having a statistical approach in biology helps but is fundamentally different than the statistics you utilize in large-scale music industry studies. Still, she's glad that she managed to combine her two passions – biology and music – and pursue both as a career. "I didn't want to give up the scientific method of doing things but also couldn't give up music, so this was a perfect choice," she concludes.