Understanding Advanced Analytics on St. Patrick’s Day
Descriptive
Straight facts compiled and examined, most effective when visualized. Tells a story of past events.
A guy at a bar says, “So then, I over leveraged my alcohol to not ratio and I heaved all over the place. Again. You can see here that I’ve done it a number of times, which is the definition of insanity.”
Implement self-service business intelligence (BI) software like Tableau or Microsoft Power BI to enable your end users with the power to interact with your data and discover insights without the aid of enterprise IT. Descriptive analytics is heuristic, offering the ability to tell a story about what happened. We can analyze data to describe both what is happening or happened through standard reporting, and ad-hoc self-service. This provides answers to how many, how much, or how often.
Diagnostic
Navigate the results to analyze problems and solutions, then take timely corrective action for better results.
Is drinking the problem? I drink too much and its affecting my relationships. Someone should slap me when I next grab a beer, or tackle me before I hit the bar.
Diagnostic analytics offers exploration and visualization of where a problem lies, and what is affected. Through alerts and metrics management, we can see what action to take and know when to take it.
Predictive
Interpreting heuristic or real-time data to anticipate future events.
I bet if I stop drinking I won’t heave any more. If I don’t, I’m going to get cirrhosis of the liver, lose my wife and end up living in a van down by the river. I ‘ve always wanted to be a spaceship fighter pilot, I can go to Space-pilot school if I stop. Maybe even start one myself.
Test product ideas on target markets, then leverage customer segmentation and predictive modeling to understand buying behaviors and propensity to purchase new products. Predictive modeling offers forecasting through the extrapolation of what will happen if elements persists or if trends continue. We can examine data and through statistical analysis discover what is possible if trends continue, or what will happen if we change x, y, z parameters.
Analyzing a Facebook feed for other posts a user might be interested in, assuming emerging buy trends based on past purchases and future needs, predicting power consumption and management, or streamlining your supply chain by betting on demand are all examples of predictive analytics.
Prescriptive
Study and learn from data to prescribe the best course of action.
Best thing that could happen is I stop drinking. In fact, everyone should stop drinking. If we just banned alcohol everything would be fine! I’m moving to Saudi Arabia.
Automate insights on complex data sets and leverage machine learning to understand and respond to real-time streaming data from sources including a grid of Internet of Things (IoT) devices. With machine learning, optimization is possible and offers answers to questions like, “what is the best achievable outcome?” or, “how can we learn from and address this automatically?”