By Scott Carey
Nov. 15, 2016
With the rise of big data and the lack of data scientists in the market, businesses are reaching out to academia more and more to help solve some of their thorniest technical problems.
In theory, it is a match made in heaven. Academics want interesting problems to solve and businesses have plenty of them when it comes to making use of their new-found reams of big data. In a world where data scientists are the scarcest of resources, partnerships between universities and businesses are helping to plug the data science skills gap.
Alice Jacques, senior data scientist, consumer insight at Channel 4 summed it up during her talk at the DataIQ conference in London last month. "Businesses have data and real problems, academics have experts and teaching capacity," she explained.
The benefits for both sides are clear to see: industry gets cheap access to data science talent and universities can tout industry experience to attract students, boost funding and help with their rankings.
Professor Patrick Wolfe from University College London (UCL) told Computerworld UK that he sees universities moving away from the "old-fashioned ivory tower model" as they look to engage with society more.
Wolfe believes that by linking arms with big business, universities have an opportunity to "apply new ideas immediately".
"If I wanted to work on a network 30 years ago there wouldn't have been much data to work on," he added. "What has happened now is there is a natural bridge."
Jacques from Channel 4 said the media company worked with two UCL PhD and four master's students on a soon-to-be released Netflix-style recommendation engine for its on-demand service All4.
Jacques told Computerworld UK that the current batch of PhD students "are doing their PhDs in recommendation systems and computational statistics, stochastic modelling and time series," which are all skills that have helped Channel 4 develop its recommendation system.
"We have six mini-experts that understand a single problem very well," she said.
Jacques urged data scientists to get out to academic conferences more. "They scare you and remind you that academics didn't stop when you left [university]," she said. "They expose you to new ideas, which you can try to fit into your business problems. It is speed dating for data science ideas."
This trend goes further than collaboration and into outright talent poaching.
"The other side is a trend of academic researchers in areas like computer science sometimes being poached away to industry," Wolfe said. "If you are excited about running algorithms at large scale you can't really do that sitting behind a university desk."
According to Wolfe, industry experience simply makes his students "more immediately marketable", but that demand for data science skills will continue to outstrip supply because "universities work on a slightly slower scale" than industry.