Big Data Meets AI: How The UK’s National Health Service Uses Analytics

Article by Clint Boulton for The Wall Street Journal, November 5, 2014

The U.K.’s National Health Service is using cloud analytics software to pluck numerical and text data on health-care facilities from spreadsheets and databases and present it in plain English on its website, NHS Choices. The endeavor, announced Oct. 28, is the latest example of how organizations are using software to manage and present large amounts of information they previously couldn’t wrangle.

The information will help England’s citizens learn more about 50,000 health care facilities and make a “more informed choice about their care,” said Cleveland Henry, delivery director, NHS Choices, in a press statement.

Managing such troves of information — commonly known as Big Data — has become a major priority for companies, said Michael Chui, a partner at McKinsey Global Institute, during a session on Big Data at the Dell World customer event Wednesday. “This is something [businesses] will have to continue to get better at,” said Mr. Chui. “It’s becoming table stakes for competition.”

NHS Choices is using Narrative Science Inc. software, called Quill, a cloud application running in Amazon Web Services. The software uses natural language generation techniques — a key element of artificial intelligence — to comb through structured data and automatically present it in story form, akin to the way a human would write a document. Kris Hammond, Narrative Science co-founder and chief scientist, said the company uses proprietary algorithms to analyze data, discern what is important and generate reports — often in under a second. The software pairs well with visualization tools that support the narrative with graphical charts.

Gleaning insights from data troves is something many companies are tackling by hiring “data scientists,” a sort of statistician and analytics expert who can glean insights from data and convey their potential meaning for a company. The challenge with data scientists is that they are “unicorns,” as McKinsey’s Mr. Chui called them, few and far between. They’re also not as fast or accurate as software, which can be an impediment for companies seeking competitive advantages, he contended.

But many companies are still trying to figure out how to use Big Data to better communicate with customers, or deliver information that helps employees make better decisions. Narrative Science is drawing interest from companies that have invested heavily in data warehouse, business intelligence and visualization technologies to corral their data, but aren’t getting a lot of value out of it, said Stuart Frankel, Narrative Science Co-founder and CEO. “The reality is setting in now,” said Mr. Frankel. “People have this data and they’re not using it.” He said the current data scientist shortage will become increasingly compounded by tremendous growth in data.

Narrative Science has 50 customers, including Forbes, Credit Suisse Group AG CSGN.VX +0.56% and American Century Investments, and is growing at a rate of over 100%, said Mr. Frankel. The company launched in 2010, initially as an academic project at Northwestern University in Evanston, Ill., where Mr. Hammond worked as a computer science professor.

The project, then called StatsMonkey, initially generated news stories from baseball game recaps, drawing the ire of journalists who feared the software threatened their jobs. It has raised $22.4 million in funding from In-Q-Tel, SAP Ventures, Battery Ventures and others.