Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
Henric Cronström,Rob Wunderlich,Miguel Angel Baeyens,
Hi I have a requirement where I need to create a pivot table with a Product field, which has to be used twice, both pivoted and non pivoted position.
The metric that I need to calculate is a venn diagram which will calculate the number of customers which use the combinations of any two products.
following is a sample of the resulting chart i need to create
Product | A | B | C | D |
A | 3 | 4 | 2 | 9 |
B | 4 | 8 | 6 | 3 |
C | 2 | 6 | 10 | 7 |
D | 9 | 3 | 7 | 4 |
The data is available is Year, Quarter, Product, Country, Customer.
I am facing a problem on how to segregate the values for a pair of particular product types. I tried a lot of ways to create a combination of two products, but I am not able to get the correct numbers.
Please help.
Thanks,
Abhinava
A classical Basket analysis...
First: How do you define a "basket"? Is it if a customer has used the two products within a Quarter and in the same Country, or is it worldwide since the beginning of time?
Once you have defined your "basket", you should define an ID for that in the script. Further, you need to create a side table using
ProductB:
Load BasketID, Product as ProductB, ItemID as ItemB_ID Resident Baskets;
Then you will get a data model like
Now you can create your pivot table using Product and ProductB as dimensions, and
Count(distinct If(ItemID<>ItemB_ID,BasketID))
as expression.
It might be possible to do all this in the pivot table using Set Analysis, Aggr and value loops, but it won't be easy...
HIC
A classical Basket analysis...
First: How do you define a "basket"? Is it if a customer has used the two products within a Quarter and in the same Country, or is it worldwide since the beginning of time?
Once you have defined your "basket", you should define an ID for that in the script. Further, you need to create a side table using
ProductB:
Load BasketID, Product as ProductB, ItemID as ItemB_ID Resident Baskets;
Then you will get a data model like
Now you can create your pivot table using Product and ProductB as dimensions, and
Count(distinct If(ItemID<>ItemB_ID,BasketID))
as expression.
It might be possible to do all this in the pivot table using Set Analysis, Aggr and value loops, but it won't be easy...
HIC
Have a look at this option
Hi HIC,
Thanks for the reply.
The "Basket" that I have to create is the count(customers+country) which use any pair of products and it is for every quarter.
So what I am thinking is to create a table which will give me the Country and Customers which use atleast two products and the combination of those every 2 products.
Am I on the right track?
A little help would be highly appreciated!
Thanks,
Abhinava
It sounds to me as if your BasketID should be Quarter + Customer + Country:
Quarter & '|' & Customer & '|' & Country as BasketID
If you use the solution sketched above, but with this BasketID, I think it will work the way you want it. But you will need to have a selection of a single quarter, otherwise the expression will count the same customer several times - once per quarter.
HIC
Awesome Henric!!
Thanks for all the help!!
Regards,
Abhinava