“Advanced Analytics” is a relatively new term in the data management and data warehousing business. “Basic” analytics is relatively a straight forward affair, mostly involved in answering ad hoc questions (that is, questions that haven’t been pre-planned) about a set of stored data. Most often these are business questions such as “How many customers have bought product X but haven’t bought product Y?” or “Did the London office sell more of product T than it forecast?” The mathematics of the question isn’t very hard… generally it is basic arithmetic logic like adding up totals or perhaps computing a single average. Nothing exceptionally sophisticated or complicated.
By contrast, advanced analytics has grown out of the scale of the data we are now dealing with. Even before there was big data, we started collecting enough data that we could start process it statistically. So advanced analytics is often the intersection of large amounts and statistics. We can start to answer some more interesting questions, predictive questions likes “Based on the last 12 months sales and social media trends, can we expect to sell more drought resistant crop seeds?” or “Is the current availability of low cost transportation going to suppress the costs of moving our heavy machinery for the next six months?” These predictions can help business optimize or improve their offerings by better understanding market and customer needs.
Advanced analytics often relies upon both complex algorithms and statistical processing to tease out trends and correlations that is not directly obvious from the data. What is often overlooked is that power of advanced analytics needs to be unleashed through availability of well curated set of data in the first place. That’s where data warehouses come in and the power of high performance data management systems come in. As we see the amount data available to us continuing to build, we begin to see the data warehouse paradigm will need to shift from basic analytics towards advanced analytics.