SKU Pearson Correlation Coefficient Analysis

JAF90

Hi all

I am trying to do a Pearson's correlation analysis to find out how likely my customers are to purchase another one of my products (SKUs) based on the overlap between them. For example, see the table below. You could argue customer C may well buy product 1000001254 as Customer A & B have it as well. In addition to volume, I have other columns but would like volume to be the key factor.

When I do the PCC calculation, it asks me for two numerical values. I have tried duplicating the same file but no luck to have 2 volume columns but its still not working properly (or dont know how I should say!). Thank you!

Customer	SKU	Volume
Customer A	1000006237	516
Customer A	1000006863	315
Customer A	1000001254	326
Customer A	1000001260	385
Customer A	1000001261	815
Customer B	1000001254	125
Customer B	1000001260	355
Customer B	1000001261	555
Customer C	1000001254	295
Customer C	1000001260	852
Customer C	1000006237	755

CharlieS

Hi @JAF90 

I don't think the Pearson Correlation tool/analysis will provide what you're looking for. I would suggest checking out the Market Basket Affinity tool. 

https://help.alteryx.com/current/designer/mb-affinity-tool 

This does not take value into account, but does show the relationship of one SKU to others based on the presence in the same transaction. I've attached an example workflow that should this in action.

20210311-MarketBasket.yxmd