Calculating success rate of Yes/No based on the difference between two dates.

Marik

Hello All,

Brand new to Alteryx & this forum so apologies in advance. I have an excel background and am trying to take advantage of some of the great tools Alteryx has to offer. I am looking to calculate a the chances of success (Yes or No) based off the amount of time someone reaches out to a client BEFORE their scheduled assignment. I'm having difficulty finding a tool that will find the difference between two dates (only needs to be as specific as # of days) and then provides estimated success based on past results. I'm still working my way through all of the new tools at my fingertips but I'm trying to produce this result faster than it will take me to learn everything. Please let me know if any additional information may be needed. Thank you in advance!

Claje

Hi,

You're looking for two tools here.

The first is a Formula.  The DateTimeDiff function is really powerful and can get you the length of days.  To configure this, I would do something like:

Datetimediff(scheduleddate,outreachdate,'days')

This will create your days between each period.

Next, you need a predictive tool.  If you're already familiar with Predictive analytics outside of Alteryx, you can skip this step, but I would highly recommend watching this "Introduction to Predictive Analytics" training as it will help you to better understand your data needs.

To quickly summarize, you are essentially calculating the probability of a binary value, which is a Binary Classification model.  Make sure to sample your data, and use the Logistic Regression tool as a starting point for this predictive analysis.  Then, once you have done so, you can use the model output from the "O" anchor of this tool, and connect it to the Score tool, where you can feed in all of your new data (where you want to predict) and see the results.

You probably want to use a Score tool against another sample of known data as well to see how the model performs and confirm that it is accurate.