Data science? What it really needs is a little imagination

"Domain knowledge can be picked up quite quickly," suggests Filtered's Chris Littlewood

One of the key talents a data scientist needs, argued Chris Littlewood, head of science at online training company Filtered, is imagination. "Domain knowledge can be picked up quite quickly. If you have got someone who's imaginative and good with the analytics, then they can easily incorporate that into the skillset," said Littlewood.

However, in a small business - such as Filtered, which has an IT department of five, with two in the role of data scientist - Littlewood added that the data scientists need to be able to see insights through to action and "some sort of change in the business".

At Filtered, he added, they didn't recruit internally from the development team, but kept them separate - "separate analyses; separate processes", said Littlewood.

However, attitudes have changed over time. He continued: "Our products are written in PHP; our data science was originally done in R. Now, our data science is being done in Python and they are moving much closer to the development team and we are trying to make sure that they use the same processes.

"The way we do it now is we try to blend the teams. It helps if you've got developers interested in data science and data scientists interested in production code."

And even the smallest of functions, they also need to be team workers too, he added. Regardless of the skills or superstar wages on offer in data science, it remains very much a team sport, argued Littlewood. "You're only going to get a good result through collaboration and sharing of ideas," he said.

Littlewood was speaking on a panel dedicated to training at Computing's Big Data & Analytics Summit 2016 last week, alongside Taj Chowdhury, information management and technology officer at the London School of Economics; Wayne Hu, head of data and analytics at Unique Digital; Anand Venu, analyst at money-transfer company Transferwise; and, Deryn Graham, senior lecturer in information systems at the University of Greenwich.

And Graham had a warning for everyone considering either setting up a data science function in their organisations, or a career in data science: she warned that, quite simply, there's not just a shortage of people to do big data and analytics, but also a shortage of people to educate and train them.

Missed Computing's Data Science and Analytics Summit 2016? Then book early for the Internet of Things Business Summit in May. Grab your place before they all go - places are FREE for qualifying end users.