Corrplot is currently used to show the result of one-by-one linear regressions. But as you know this only gives an impression which variables are related. In a mix however and using multiple regression, different variables could all together impact a certain response variable. It is this result set that I want to show using corrplot.
In an offline version I already implemented the possibility to show the result of multiple regressions sets in a similar way. Predictors will be shown on the vertical axis and one or more response variables on the horizontal axis. Variables that are excluded in the multiple regression will be crossed out. Variable with no influence will be blanked out. For this type of diagram, the number of items on each side will differ by default.
As an MVP only "Circle" will be implemented using the mtcars data set.
Example:
I intend to add this functionality in the coming months, time permitted.
Corrplot is currently used to show the result of one-by-one linear regressions. But as you know this only gives an impression which variables are related. In a mix however and using multiple regression, different variables could all together impact a certain response variable. It is this result set that I want to show using corrplot. In an offline version I already implemented the possibility to show the result of multiple regressions sets in a similar way. Predictors will be shown on the vertical axis and one or more response variables on the horizontal axis. Variables that are excluded in the multiple regression will be crossed out. Variable with no influence will be blanked out. For this type of diagram, the number of items on each side will differ by default. As an MVP only "Circle" will be implemented using the mtcars data set. Example:
I intend to add this functionality in the coming months, time permitted.