I am trying to use the Linear Regression, but I am not sure how to select the fields for each category. My goal is to determine if there is a correlation between % discount (x) and sales quantity (y). I have a sales table with 5 years of data.
Can anyone provide an easy example so I get the idea on how to use the action?
Thanks in advance!
The typical use case for linear regression is understanding the linear trend of a sequence of numbers. For instance, to understand if values are growing or declining in general over the specified period of time.
I don't think it's the right tool to understand correlation.
PS. As a side note, understanding the correlation between a discount and sales volumes is a very non-trivial task because usually there are multiple factors (besides the discount) that affect sales. I'm not sure if you get a meaningful answer if you calculate the correlation just between discount and sales, unless you have sales data for a reliable control group of customers that we similarly active but didn't get discounts.
From my experience, if, without calculating a correlation, it's not obvious whether a discount led to an increase in sales or not, then most probably the effect of the discount was too minor or insignificant.
Hi! thanks for the quick response.
Exactly, I agree with your arguments. We were just trying to get an example on how to use the action since it was the first time I saw it.
As you mentioned, we have a control group without discounts for each sales channel (they have different purchasing patterns) to compare with the ones having the discounts in the same period of time. We are exploring our data at this point, so for these first steps we always use Easymorph to get a quick sense of what we are trying to get.
I think we misunderstood the functionality, and we couldn't find enough documentation to understand the logic behind it (basically, we thought it was "y = f(x)"), so thank you for explaining it.