Welcome back! On this weeks episode of Data Science Wednesaday, Decisive Data's Lead Data Scientist Tessa Jones takes us through a basic understanding of what prescriptive analytics is. Check back next Wednesday for an all new DSW video!
What is Prescriptive Analytics?
Today, we're going to be talking about prescriptive analytics, which is really the last set of analytics within the spectra of analytics that we have covered.
Reviewing Descriptive, Diagnostic, and Predictive Analytics
Descriptive analytics is really the baseline of all analytics in general, as this really tackles the question of: what's going on in my business? Then, we move into diagnostic analytics, which gets us into why things are happening, why is my revenue going up or down, things like that.
Diagnostic analytics really help us to understand how to predict things—which leads us right into predictive analytics. Predictive analytics provides a model that's going to tell us what's going to happen in the future. So, now what? You know what's going to happen in the future.
The beautiful thing about prescriptive analytics is it tells you what to do with that information, and it gives you an action that you can run with to drive your business.
Example of Prescriptive Analytics
For example, let's pretend that you are a grocery store owner, and you have a good baseline of descriptive analytics, and some nice dashboards to help you understand what's going on in your business. You also have information about how different products are selling over time, things of that nature.
Then, you want to go into diagnostic analytics, where you're looking at why things are happening and what causes revenue to go up or down?
All of that really supports the building of a predictive model.
This is an example of the output of a predictive model. We have two products here. One is blue, and one is red. And the blue product here, you can see that it's very seasonal, right? Like, the blue product sells a lot more during the summer, and the red product sells a lot more during the winter. And vertical line indicates what's going on today.
The predictive model shows that blue is going to go way up, and the sales in red is going to go way down. So, now what? What do we do with that information?
That's where prescriptive analytics comes in. We integrate that with the descriptive information we have, which, down here, this is telling us what is on hand. We can see that the blue product, the inventory in the store, is going down and we're about to run out, whereas the red products, we have an abundance of those.
If we take this information and integrate it into the business process, we can say, "Well, for cereal on week 36 in store 10, we predict that we're going to sell 56 units of cereal, but we know we only have 40 on hand." We also know that this particular business doesn't want to ever run out of products because, to them, that means they're losing revenue. So, we want a buffer of that, so we decide we're going to ship 18.
Conclusion
If this is really a good prescriptive model, then this is seamlessly integrated into the business so that these actions are just happening. It frees you up to make all kinds of other business decisions that are very important and more insightful.
To recap, prescriptive analytics is the suite of analytics that gives you actionable things to do with the data that you have. That's prescriptive analytics, and thanks for joining.
Posted by Gage Peake
