Wanted: Apps With Data Science Baked In
February 6, 2013-
By Jeff Bertolucci
Many organizations launching big data platforms are finding that data scientists are in short supply. How bad is it? The McKinsey Global Institute predicts the United States alone could face a shortage of 140,000 to 190,000 data analytics experts just five years from now.
Granted, the rarity of the elusive big data guru might be a short-term problem. More business schools, for instance, are offering analytics coursesthat give students a basic set of data science skills, a trend that might enable highly trained data analysts to focus on more complex matters.
Still, democratizing big data is a worthy goal. But how do you go about it? According to PROS, Inc., a pricing and revenue management software company, the answer is to create apps with the data science cooked in.
“Data scientists are hard to find, there’s no question,” said Pros chief marketing officer Tim Girgenti in a phone interview with InformationWeek. One way to close that gap, he noted, is with “an application that has the data science embedded in it.”
“In the future, you’ll buy business-oriented applications that understand your industry segment,” said Pros CEO Andres Reiner. “And you’ll have already configured the data science capabilities that most matter in your business.”
The idea of big data apps is growing increasingly popular, and the Pros executives aren’t the first to pitch custom applications as a potential solution to the data scientist shortage.
Laura Teller, chief strategy officer for predictive analytics firm Opera Solutions told InformationWeek in January that big data applications could help automate many data scientist tasks. These apps could help with simple, basic analytics, enabling non-techies to make data-driven decisions without consulting the staff big-data guru.
More sophisticated analysis, however, would likely require experts trained in a variety of technical disciplines, including computer science, analytics, math, modeling and statistics. Pros sells big data applications to help companies drive sales growth. The company started 27 years ago with a focus in the airline industry, and has since expanded to other industries such as manufacturing, distribution, services and travel.
For a big data app to succeed, it must meet the needs of a specific industry and address real-world problems, the Pros executives said.
“Take a wide variety of data, whether it’s structured or unstructured, pull it together, and apply data science to predict a business outcome,” said Girgenti. And then, “understand what the likelihood of the future is going to be, and prescribe actions around it.”
“Companies are now expanding beyond relational databases, and are leveraging technologies that offer real-time processing and analysis,” said Reiner.
The move from “disconnected to connected data” is another emerging development, and Reiner used an automotive analogy to explain the trend. “Cars have become much more sophisticated over time,” he said. “We started with collision detection, for example, with sensors that beep when you’re approaching an object.”
But newer, more advanced car safety features combine multiple components. If sensors determine other vehicles are too close, for instance, the vehicle can automatically activate the brakes to avoid an accident. And Google’s driverless car takes things a step further by adding GPS to the mix to drive the vehicle automatically.
“These components are starting to be used in conjunction in the car. They’re not decoupled, they’re connected,” said Reiner. “In the future, data must be connected and processed in real time too. And it should give simple prescriptive action.”