Channel Daily News

How to overcome the domain expertise barrier to Big Data

A lack of domain expertise is the largest barrier to companies accessing big data, according to analytics software developer SAS Canada.

Mike Luke, data management practice lead of SAS Canada made the remark at a panel discussion in downtown Toronto today hosted jointly by the Information Technology Association of Canada (ITAC) and market research firm IDC.

The panel also included David Corrigan, director of marketing for big data, integration and governance at IBM, Gagan Mehra, head of product marketing at Software AG, Gokula Mishra, vice president of advanced analytics and big data at Oracle, and David Senf, vice president of infrastructure solutions research at IDC Canada.

They found that only 16 per cent of organizations have staff in place that are able to work on big data projects.  In addition to looking into hiring more data scientists, and those skilled in statistics, math, predictive analytics or machine learning, organizations may want to rethink their strategiy, the panel said.

“Every role that you already have needs to be reskilled based on big data,” said Corrigan.  “Big data forces us into this area of relative decisions.  It’s not a black and white; is the data perfect, is it secure? But rather is it secure enough for the things that I’m doing? Is the quality good enough? It’s actually incredibly difficult to make those relative decisions and tradeoffs.”

However, it may not be necessary to lay off your entire staff just yet.

For one, better tools are coming.

Mehra explained that the solutions available for analysing data today are already more accessible to those without specialized backgrounds, and that they would become even simpler in the future.

Furthermore, not all companies may have a use for big data.  While it seems counterintuitive, the panel emphasized the importance of determining “why” a company wants to use big data.

“If you have a sampling issue where you’re not able to look at all your data to solve a challenge, you have a big data issue, if you have a challenge around data processing speeds and want to make it a lot faster, you might be having a big data challenge,” said Mehra.  “There’s a variety of things in your environment where you could be having a big data challenge, but before you go out there and adopt the big data technology, you need to evaluate what exactly is happening in your environment and figure out a business case.”