Counting the Occurrence of Items Frequently Found Together Within a Group

[WJdata]

Hi guys, I am trying to identify and count the occurrence of products that are commonly bought together. You could think of these products as complementary goods, like a cell phone and a cell phone case.

This data set contains over 1,000 products, and I am interested to see which products tend to be purchased with one another. Orders can range in size from 1 item to 20 items. The goal is to find the pairs of products that commonly occur in customer orders and identify the top 25% of pairs by frequency.

My first step is to group by order number to see which products (SKUs) are purchased in a set. Where I could use some help is to find the common pairs of complementary goods within each order.

I attached the data set if you want to give it a look. I would love any guidance you could provide or any insight into ways to solve this challenge.

Thanks!

[fmvizcaino · Accepted solution]

Hi @WJdata ,

I'm attaching a solution where I'm using a join + a filter tool to create all possible combinations of 2 itens from the same order.

Take a look and let me know if you need further assistance.

Best,

Fernando Vizcaino

[WJdata]

This is perfect @fmvizcaino! What an ingenious way to think about it, thank you! 

If I wanted to continue the analysis with say 3, 4, 5...+ products would I continue the pattern by adding a SKU ID 3, SKU ID 4, SKU ID 5, then adding more filter tools until all 5 > and < scenarios are accounted for?

Thanks!

[fmvizcaino]

Yes! For each join added, you need to create one additional condition in your filter.

Best,

Fernando Vizcaino

[ynilabh]

Hi Please help me on this tool , I need to create 3 combinations but unable to do so

[ynilabh]

Ho to create a new join for 3 ,4,5 combinations ,please help