It's been called the "new Latin" for data science students, and for those who get their degree it's considered the "sexiest job of the 21st century."
David Wagner had a question for his students. Standing in front of one of the fastest-growing classes at UC Berkeley, the professor explained lines of code used to construct a graph comparing pressure in footballs used by the New England Patriots during the 2015 AFC Championship - the subject of a hashtag-ready scandal known as "Deflategate."
"You're the consultant reporting back to the NFL on what you've found. Do you buy their story?" Wagner said, prompting boos from the lecture hall as the students - like the NFL - concluded that the Patriots had deflated their footballs.
In spring 2016, UC Berkeley's first foundations of data science course attracted around 300 students. This semester, about 1,300 have enrolled, and as of this month, most eligible undergraduates can choose data science as their major.
Dubbed the "sexiest job of the 21st century" by Harvard Business Review, data science is most easily described as using the modeling and analysis skills of statistics combined with the programming and machine learning tools of computer science to find patterns in data and extract insights. But sexiness hasn't lured enough people into the field, especially around Silicon Valley.
The workplace social network LinkedIn reports a shortage of more than 36,000 people with data science skills in the Bay Area, a 33 percent increase from the start of the year - and a stark contrast to 2015 when the nation as a whole had a surplus of such skills. (Those numbers don't mean that would-be data scientists were necessarily unemployed, LinkedIn said; it could indicate that they were "underemployed" in the sense that they worked jobs that didn't use their data science skills.)
Big data are transforming every industry, and there's money to be made in understanding how to manipulate them. It explains why universities are rushing to add data science programs and why UC Berkeley students from a hundred different majors have enrolled in Wagner's course this semester.
At Macy's, data science helps the aging retailer follow new customer preferences and drive sales growth. UCSF is developing a tool that would let computers scrape millions of electronic health records to detect patterns in disease. Allbirds, a San Francisco startup that sells shoes made from wool and eucalyptus fiber, is expanding its two-person data team to help decide on the quantities it makes of different shoe styles and colors, as well as how much can be sold in various regions.
Mark Levy, head of the employee experience at Allbirds, said by email that the company gets a "constant stream" of applicants for data science openings, but nonetheless, "because we're a lean team, we have to do a lot with a little and it can be difficult to find candidates with the right technical skills to do it all."
Adam Bennett, Silicon Valley director of technical recruiting for full-time hires at IT staffing company Robert Half Technology, said his firm has been inundated with requests from small and midsize clients seeking their first data science hires this year.
A person who is "classically educated" at a top-ranked university and has a master's or doctoral degree in mathematics, statistics, economics or physics is likely to be recruited before graduation, he said. After they gain real-world experience, such people may have "anywhere from three to five offers" when they're ready to move on.
"All it takes is for them to apply to one other job, and before you know it, they're off to the races," Bennett said.
The median base salary for a data scientist in San Francisco is $189,658, according to research from Robert Half Technology. The top 5 percent of candidates earn $246,750 a year and up - on top of other types of compensation, such as stock options, restricted stock units or annual bonuses.
Established firms such as Facebook and Google can pay better, so small and midsize companies look for other ways to compete.
Reddit, the social aggregator site, tries to sell job candidates on the impact they will have, rather than just salary, according to Lin Huang, the company's director of data science and engineering. His team uses data to decide what features should be built next to increase user engagement - basically, designing the app to be even more addictive.
In interviews with candidates, Huang looks for a sense of curiosity. Data scientists are "always hungry to learn more," he said.
The demand from outside the tech and finance industries for people with data science skills has squeezed the local hiring pool, according to Guy Berger, the chief economist at LinkedIn. And plenty of data scientists prefer to put their skills to work outside the Bay Area, with its soaring cost of living.
Reddit encourages employees to brush up their data skills by taking online classes, according to Huang.
Airbnb runs its own program, called Data University, to help employees become more data-literate. It has about 120 data scientists at Airbnb globally, according to a company spokesman, and has 15 additional openings in data science and analytics at its San Francisco headquarters. Krist Wongsuphasawat, a data visualization engineer at the company, said that "there was no data science" when he began his postgraduate studies in computer science in 2007. Now, everything has changed.
"Data science is the new Latin for university students," said Wagner, who was part of the team that developed the data science course at UC Berkeley's Department of Electrical Engineering and Computer Sciences. There was a time, not too long ago, that to be a college-educated person, "you had to learn Latin because that was the language of scholarly study," Wagner said. "Now, data is the coin of the realm."