3 Replies Latest reply on Dec 6, 2017 4:24 PM by swaroop.gantela

    Market Basket Analysis without "Self-Join"

    Lawrence Steffes

      Hello fellow Tableau forum experts! [This is my first post FYI]

       

      I need some help - I'm attempting to conduct a market basket analysis on my boolean data set without utilizing the famous "join method''. In order to be courteous to those [less experienced colleagues] I will share the information with in the future (via desktop) I would prefer to avoid the self-join that seems to be common. I would also like to visualize these relationships in a heat map format as shown. I have partially constructed the view, but validation shows some of the relationships described are not accurate. For instance, look at the relationship between rows and columns AAA & BBB. Could this be an issue of directionality of the relationship? I'm also trying to understand why my calculated field (c_affinity) lacks a field that is clearly input in the calculation (DDD). I really appreciate you input in helping me to better understand this complex view and make it work, if possible. Thanks in advance!