Skip main navigation

What's the difference between predictive and prescriptive analytics?

Predictive and prescriptive analytics are different, and related. Find out more in this step.

Top tip: There’s a lot of information in this step that helps to give context to the example seen in the Helios video above. We recommend that you do the reading first, then watch the video when you reach the prompt.

Title illustration, 3 fields. Top: "Predictive analytics", "What is likely to happen next?", "Foresight". Bottom: "Prescriptive analytics", "Based on modeling, what could we do as a result?", "Amplified intelligence." Side: "Looking into the future".

Although predictive and prescriptive analytics are different, it’s useful to discuss the ways in which predictive and prescriptive analytics are related to each other.

While predictive analytics uses historical data and statistical modeling to predict likely outcomes, prescriptive analytics takes it a step further. It allows organizations to ask, “If these are some of the things that are likely to happen in the future, what are some of the proactive strategic decisions we can take and what are the likely outcomes of those decisions?”

While predictive analytics makes the case for what might happen in the future, prescriptive analysis provides decision-makers and internal teams with optimal actions to reach their future goals.

Chart, 2 columns, 4 rows. Column 1: predictive analytics. 2: prescriptive analytics. Rows: 4 capabilities with a yes/no answer. Eg row 1: “forecasts and predicts future outcomes”. For predictive analytics: “yes”. For prescriptive analytics, “no”.

Machine learning & AI

Before we get into how predictive and prescriptive analytics use AI and ML, let’s remind ourselves of these terms:

Machine learning (ML) gets machines to do new tasks without being explicitly programmed to do so. This is done by providing data to machines and allowing them to use that data to learn for themselves. ML gathers, sorts, prioritizes, and finds patterns in data to reliably forecast trends and behaviors in order to understand and predict future outcomes.

Artificial intelligence (AI) uses computing power to perform certain tasks that traditionally require human intelligence. AI uses ML, algorithms and neural networks to analyze, classify, and draw meaning from data. Unlike regular programming, which works within the constraints of its programmed scenarios, AI is able to explore, learn, and improve on its own.

AI uses ML to generalize and learn from past experience so it can discover meaning and gain the ability to reason.

Prescriptive analytics is usually accepted as an extension of predictive analytics. In predictive analytics, data scientists find patterns and make predictions about future events, in order to gain insights from this data. Insights obtained through predictive analytics can then be used further within prescriptive analytics to drive actions based on predictive insights.

So, even though you can start a data analytics investigation with prescriptive analytics, the process you would go through would incorporate some of the same processes used by predictive analytics that we covered in the previous lesson.

In short, prescriptive analytics takes the results of predictive analytics that were found using ML and applies AI to take the next step.

Let’s use the example of autonomous cars to help illustrate the relationship between predictive and prescriptive analytics.

Autonomous cars should be dangerous. How can you expect cars to drive themselves? Data analytics is the answer. Autonomous cars work in part by using predictive and prescriptive analytics to analyze every single possible outcome and make decisions about the best course of action.

For example, when determining the distance in which to make a safe stop at a stop sign using sensors, we can take the data from thousands of recorded instances where cars have made a stop at a stop sign and calculate the optimum distance at which to start applying the brakes.

Predictive analytics

Using machine learning, predictive analytics calculates and predicts the distance from the stop sign that the car should stop. What predictive analytics can’t do is provide the car with the instructions it needs to stop safely or execute those actions. For that we need prescriptive analytics using AI.

Prescriptive analytics

Prescriptive analytics uses AI that is trained using enormous sets of data that are too large to be analyzed and sorted by humans. It provides the autonomous car with the best actions and processes the car needs to perform. It predicts the possible outcomes of either taking those actions, or not taking them.

In the case of the autonomous cars, prescriptive analytics is not able to perform the prescribed actions on its own. The results of the analysis are processed which, allows the car to perform those actions and carry them through without the need for human input.

There’s so much more that happens behind the scenes with machine learning and AI than what we’ve covered here. For our purposes it’s good to understand the relationship between machine learning and AI as we start to understand future facing analytics.

Okay, now we’ve covered what prescriptive analytics is, what it does, and how it differs from predictive analytics, let’s catch up with the team at bike company Helios.

Watch the video above

The last time we checked in on Helios, they had used predictive analytics to work out that a partnership with Rollic (the e-wheel company) would be the best option for them.

Taking these findings, Helios proposed a potential partnership to Henny Van de Berg, CEO and founder of Rollic. Both organizations went away and had their data teams work out an optimal strategy for how they could combine aspects of their operations.

Predictive analytics told both Helios and Rollic that a partnership was a good idea. Using the data-sets and models from both companies, prescriptive analytics will give them some optimized suggestions for the specific steps needed to get there.

Right now, Caroline, COO of Helios, and Joel, the company’s data analytics consultant, are having another meeting with Henny to run through their latest findings…

Key takeaways

  • Rollic are keen for a partnership with Helios as it will give them the opportunity to capture market growth effectively and impactfully
  • Rollic’s own data analytics projects had also identified San Francisco and Amsterdam as initial target cities
  • Helios and Rollic are going to combine their data-sets so that they can better understand the potential benefits of their proposed partnership
  • Both parties also hope that prescriptive analytics can suggest the steps they need to take to achieve their shared aims

We’ll check in with Helios and Rollic later, to see how their plans are progressing. But it looks like this partnership is going to benefit both companies.

Speaking of benefits, let’s have a look at the benefits of prescriptive analytics for organizations and individuals…

This article is from the free online

A Beginner’s Guide to Data Analytics

Created by
FutureLearn - Learning For Life

Reach your personal and professional goals

Unlock access to hundreds of expert online courses and degrees from top universities and educators to gain accredited qualifications and professional CV-building certificates.

Join over 18 million learners to launch, switch or build upon your career, all at your own pace, across a wide range of topic areas.

Start Learning now