Summary

We have created a fast, stand-alone Key Driver Analysis Tool that performs both Shapley and Kruskal analyses, in addition to Ridge Regression. The results are typically produced in seconds, as a bar-graph and .csv file. Even with 20 independent variables, the tool returns results in only a couple of minutes.

key driver analysis tool

NEW! Super-Fast Online Relative Importance/Key Driver Analysis Tool FREE Two-Week Trial

Read More

The Issue

Shapley Value Coefficients and Kruskal’s Relative Importance Weights are two of the most advanced techniques used to accomplish Key Driver Analysis. However, their use is often limited by the expertise and software required, and the time taken to run both analyses when many variables are considered. We wished to create an easy to use, time-efficient program that conducts both forms of analyses simultaneously.

Our solution

We wrote, then implemented in Python, an efficient recursive algorithm that simultaneously creates both Shapley and Kruskal weights, and exports the results as required. The program also allows the user to simultaneously conduct Ridge Regression on the same data-set and export any relevant output.

the result

The program allows end users to conduct various Key Driver Analysis techniques without requiring multiple software packages and/or specialist knowledge. It can be easily customised to read all common data files.
To read about and download JumpData’s Key Driver Analysis tool, click here.

Are you interested in reading more about our projects? Choose a category from the buttons below:

Other projects in this category: