When it comes to making the most of the potential offered by big data analytics,Read More
What skills do your data science team need?
When it comes to making the most of the potential offered by big data analytics, it will be vital to have the right staff on hand to manage the solutions and interpret the results.
But this could pose a challenge for many firms, as competition for the best talent is fierce. Indeed, it was noted last month by the Wall Street Journal that someone with ‘data science’ in their job title on LinkedIn could get up to 100 recruiter emails a day, according to Josh Sullivan from consulting firm Booz Allen Hamilton.
With this in mind, it may be tempting for many businesses to look to their own ranks when trying to fill a data science position, rather than competing for the scarce resources in the recruitment market. But what will firms need to look for when aiming to give staff the skills they need to help create an effective big data analytics strategy?
Kevin Lyons, senior vice-president of analytics for digital marketing data management platform vendor, told CITEworld that in order to achieve this, companies first need to have a clear idea of what they want their data science team to achieve.
“The first step is to define a clear business goal, or at least one the company is working toward,” he said. “If you can’t identify it, there’s no way you’ll be able to achieve it.”
This may vary from firm to firm. For instance, tech-based enterprises such as Google and Facebook need analytics that will help them learn about consumers and predict behaviour – for which strong mathematical and computational skills are a must.
However, if the goal is to produce analytics to help humans make product or operational decisions, effective ‘soft’ skills will be necessary.
Mr Lyons explained that every data analytics project can essentially be broken down into four components. These include understanding the business need, gathering and preparing the data, undertaking the modelling and translating the outcomes into operational results.
These will require very different skillsets, so it is important the people working on the project have a mix of talents. Not everyone needs to be a qualified data scientist with a strong understanding of the specific tools a business will use to study its data, as good communicators and statisticians are also highly important.
Claudia Perlich, chief scientist at Dstillery, a marketing company that analyses web browsing data to help brands target ads, said at least one person on the team needs to be an effective communicator, in order to explain the project and detail to other departments what they are trying to achieve.
She described this as “somebody who can sit down with the CTO or CMO or CEO and have a good enough understanding of the business problems to help frame what role and what specific task data science should work on.”
One of the big questions for businesses – particularly if they are trying to decide whether to bring in outside experts or promote and train from within, is finding the balance between a good knowledge of the business and a strong analytical understanding.
For many companies – particularly those with highly complex operations in fast-moving sectors – the value of a thorough grounding in the enterprise will be invaluable. However, if the right external candidate should come along, Ms Perlich said businesses should not be overly concerned about the time it will take to impart this to a new hire.
“If they’re smart enough to be data scientists, trust me, they can learn about your industry in a month or so. Don’t worry about it,” she said.”