Data science can enable you to turn the vast quantities of data sloshing around today’s enterprise into actionable insights and to make predictions. It uses scientific methods and technology to enable you to interrogate the data you collect to accomplish a set of business goals.
What is it?
Data science combines methods, processing, knowledge and tools to extract meaning from data. It may involve mathematics, statistics and computer science. If your business aims to create a competitive edge through data mastery, you’ll need data scientists.
Data science should be seen as distinct from the historical idea of a ‘data analysis’. Both can be used to analyze historical data to draw inferences. But what sets them apart is the use tools and techniques to model the increasing complexity and uncertainty in the world. Data analysts have usually relied on rules-based approaches that struggle with complexity and uncertainty.
A good example of how data science differs from previous data analytic approaches is the way it can deal with bad data. In old data analytic paradigms, if you had bad data, it’s rules-based approach meant you got bad results. But data science can cope with noisy data, it’s tolerant of bad data.
What’s in for you?
Data science will enable you to improve your understanding of your business operations by putting in place the foundations for detailed analytics. But what sets data science apart from traditional ‘business intelligence’ type activities is the use of modern machine learning and deep learning tools that enable organizations to accomplish their business goals.
If your enterprise is able to interrogate data — whether that’s traditional, structured operational data or messier, more complex ’big data’ — and draw out and exploit business critical insights, you have a significant and potentially sustainable source of competitive advantage.
What are the trade offs?
Data science is a discipline. You can’t simply buy data science; even employing data scientists isn’t enough. You need to be capable of engaging highly skilled people and implementing systems and processes that will get the best out of them.
But that comes at a cost. Skilled practitioners — data scientists — might well have PhDs; they’ll need to have studied extensively to gain sufficient proficiency in maths, statistics and computer science — and that makes them difficult to attract and expensive to retain.