If Goldilocks had Advanced Analytics

There is a lot of talk about advanced analytics these days. The onslaught of marketing around predictive analytics and machine learning is creating a virtual ‘gold rush’ of IT organizations scrambling to understand how they can get that edge promised by advanced analytics.

The problem is that most people don’t really understand what advanced analytics is or specifically how the technology actually works in order to design solutions to take advantage of it. That fact is a potentially huge barrier of entry for IT organizations. In many ways advanced analytics is alien technology, foreign in practice to even those who often work closest to technology; DBAs and BI teams. Machine learning, semantic & sentiment analysis, and neural networks are just a few of the various techniques available within the realm of advanced analytics and they all require time to understand how to properly apply them.

With that in mind, let’s take a moment to delineate between business intelligence (BI) and advanced analytics.  A simple explanation is that BI focuses on the past, what happened, how many or how often, whereas advanced analytics focuses on the future, why something is happening and what will happen next. BI typically focuses on reports and queries. Advanced analytics is about predicting the next best action, optimization and/or correlation. Essentially, where BI is the processes of retrospective analytics, advanced analytics is about what predictive analysis.

At the core of both BI and advanced analytics we have some sort of data storage infrastructure, whether SQL or NoSQL. Which infrastructure you use is dependent on what type of results you’re looking for and how you create those results. Building the BI or advanced analytics solution is dependent on many factors. Who is consuming the information? How advanced of a user are they? How much visualization do you need? What results are you looking for? Sophistication of solution has a broad range, with Excel at one end of the scale and Hadoop with HDInsights at the other end. In the middle of this continuum sits SQL Server and R.

There are many components of an advanced analytics solution including data ingestion and pipeline, machine learning and analysis (or other techniques) and dashboards and visualization. Within each of these categories you can get as complex as you wish. For example, with dashboards and visualization you may just want to display real-time results of data as it’s being streamed in, or you may want a visualization that has multiple dimensions, all of which you can drill down into as you click on various sections of the visualization.

As I’ve already mentioned, the key difference is that BI focuses on what has already happened.  Advanced analytics takes the next step and provides us a way of focusing on what to do next.  Focusing on the future, instead of the past. Business have historically used BI to understand their business. That’s still relevant today. But organizations that take it upon themselves to use the data to determine the best ‘go-forward’ path will find themselves with an advantage over companies who rely solely on BI.

We have an advanced analytics resource page with a lot of additional information. Check it out to discover many of the ways Advanced Analytics can help your organization!

About the Author:

Leave A Comment