Predictive and prescriptive

- I'm obsessed with tracking and rating books I read with a website called Goodreads. It helps me keep track of everything I've read so that I can identify related books that I'm likely to enjoy or find fulfilling. Classifying the books I've highly rated in order to identify new books I'll probably like is a great example of using predictive analytics to improve my experiences. Similar to Goodreads, predictive and prescriptive experience analytics uses historical information to help improve learning experiences. Predictive learning experience analytics uses data that you've collected to predict how experiences will be interacted with. We've seen predictive analytics concepts applied to learning experiences in a variety of ways. For our first example, we'll look at how to apply predictive analytics to Souder Manufacturing's course assessments. At Souder Manufacturing, we've been tasked with training location managers on a new production management software. In reviewing the results of a particular assessment, we see two important trends. Of 40 questions in the assessment, five are highly correlated with success, and people who finish faster tend to be the better performers. Using a correlation model to identify the first trend is an example of trying to predict the outcome of the experience sooner than previously available. Trying to change for better or worse a future state we've predicted is prescriptive analytics. In the context of our Souder example, because we now know certain questions are associated to higher or lower levels of success, we may be able to alter the learning path or assessment to provide a more efficient experience. Making suggested changes based on our predictions is an example of applying prescriptive analytics. If we strip away unnecessary elements, we can streamline the course and assessment and hopefully improve overall performance as well. If we take this concept of modifying courses or assessments and scale it across different experiences, you may recognize it as adaptive learning. Adaptive learning is using predictive analytics to determine what you think is going to happen and then changing the experience in response to those predictions. This makes the experience more efficient and helps the learner learn faster. In the two examples we've discussed, one is more approachable for an individual to accomplish, while the example of scalable adaptive learning likely requires third-party software. There are lots of available tools to help you along your analytics journey but you might be surprised how easy it is to get started with what's already available to you. Tools that you might already be using, like Excel, can help you identify correlations. I encourage you to think of a set of learning experience data you have that might be interesting to explore, whether they're test results, learner surveys, or utilization metrics. How might you modify the experience in the future based on credible correlations you find?