Big Data is the next Big Thing – but before throwing fistfuls of money at big data initiatives, organizations need to make some decisions: Why do analytics – what’s the business case? What tools should we use? Should we build a team in-house, or hire an outside vendor?
One large deciding factor (and answer to the last question) is the seeming lack of data scientists, as demand for data scientists continues to grow exponentially. If your organization is serious about crunching data, however, you’ll need to either find the talent and / or develop the talent.
As InfoWorld advises, the first step is clarify why you’re assembling a data science team. Without a goal – a business case for data analytics, there’s no way you’ll gauge any success.
One of the key members of your team will also have to have a foot in both camps: they will need some “soft skills” but have some technical prowess. They will need to communicate with the C-level team members, be able to explain both the data and the implications on business – but also know some scripting (e.g., Perl or Python), be able to crunch numbers, manipulate the data and have a practical understanding of statistics.
Of course, the team should also include rock-solid data science skills: members with a background in computer science and modelling statistics; with a familiarity of procedural languages such as Java or C, Unix and Linux.
InfoWorld cites Kevin Lyons, senior vice president of analytics for digital marketing data management platform vendor eXelate, who lays out the four components of a data project: understanding the business need; gathering and preparing the data; doing the modelling; and operationalizing the outcome.
If finding the talent with the right data science background who also understands the ins and outs of your business seems to be an impossible combination – don’t worry. Chances are if you find strong data science talent, they can learn your industry. The data comes first.