Big data pillars you need to get right as individual, business

Talent in the form of data scientist is non-negotiable. FILE photo | nmg
Talent in the form of data scientist is non-negotiable. FILE photo | nmg 

Often, the challenge is distilling technology for the critical parties, namely business executives and the consumer to drive purchase and adoption.

Big data as an operative term has been around since the ‘90s, only recently attaining its buzzword status. It has always been there, we just did not give it right focus.

Two factors triggered this shift; first, better tools and infrastructure and second, increased competitiveness that has everyone looking for that extra edge. Simply put, big data is the sum total of all the pieces of information we and the businesses generate daily.

There are four core features inherent to the classification of big data. Volume looks at the size of data from the bit to the exabyte and beyond.

Velocity is all about speed to insight; the collection, storage, processing and visualisation. Variety covers the fact that we have moved from the age of cookie cutter information, accounting for differentiated sources, formats and structure.

Last is veracity that acknowledges the uncertainty of data; bias, noise and abnormality, an obvious challenge when variety and volume are thrown in same pot.

Businesses of any size can benefit, but first, a strategy that informs what is to be achieved as outputs must be in place.

The business must also hire for the role of data scientist or data wrangler.

Challenges do exist though. There is no single analytics method that is good for everything so the department, team or individual would need to be conversant with multiple tools and methods of extraction, analysis and presentation.

Also, the breadth of the art and science, means that it is very difficult to cover all the bases or assure maximisation of data generated; a constant process of learning and experimentation.

Big data should not be synonymous with big budgets and there are no hard and fast rules here.

You can go the big vendor way or roll out custom solutions based on open source.

In essence you should spend as much as you need to stay ahead or get ahead; and that cost would vary greatly from one company to another, depending on the stage and vertical of operation.