Main image of article 3 Considerations Before Buying Into Big Data
Another day, another Big Data platform designed to democratize analytics. At Microsoft’s Worldwide Partner Conference this week in Orlando, the company introduced the Cortana Analytics Suite, which—like so many platforms before it—is designed to bring the power of predictive analytics to as many people as possible. It’s cloud- and subscription-based; in addition to processing structured data, the software is capable of digesting unstructured information from sensors and devices, making it a potentially useful tool for any company building products for the Internet of Things. The Dartmouth-Hitchcock Health System in New England is reportedly using the Cortana Analytics Suite to development treatment plans for patients, a test case that sounds remarkably similar to what IBM is doing with Watson, its own artificial-intelligence platform, in the oncology field. Check out the latest analytics jobs. Microsoft plans on integrating the Cortana personal assistant with the analytics suite, in a bid to make working with the software a little more human-friendly. But with so many platforms on the market already offering (or at least promising) actionable analytics, can Microsoft carve out a space for itself? Considering the platform won’t hit the market until later this year, it could be some time before anyone can judge Microsoft’s success (or lack thereof) in that endeavor. In the meantime, any tech pro considering whether to use a data-analytics platform should consider the following points:
  • Big Data Won’t Solve Everything: When it comes to analytics, lots of software vendors promise the world. But while most analytics packages are very good at sussing out relationships between objects or figuring out trends, they can’t provide all the answers. The solutions to lots of real-world problems still hinge on human intuition and experience.
  • Correlation Isn’t Causation: Just because your expensive analytics package produced a statistically-significant correlation doesn’t mean there’s an actual connection between two trends. Questioning the results of analytics software—no matter how sophisticated—is key.
  • Big Data Needs Lots of Data: Some businesses generate reams and reams of structured and unstructured data. Others (especially small ones) barely produce enough to fill two columns on an Excel sheet. If your business falls into the latter category, you may not need an ultra-expensive analytics package.
While there’s a lot of pressure to go big (so to speak) with data analytics, the decision to go with a particular platform shouldn’t be taken lightly.