HI guys,
Been trying to create a market basket analysis using splunk. Simply, I would like to build a pivot that would show me what other items were bought when Product A was purchased.
I have a CSV with two columns, Transaction IDs and Products names. If a transaction had more than 1 item purchased the id will be listed with as many unique items as possible.
Example
ID-32 Product A
ID 32 Product B
ID 33 Product G
ID 33 Product B
ID 34 Product A
ID 34 Product J
etc.
I am sure Splunk can do it. I just can't wrap my brain around it right now.
Thank you.
this search string doing the work for the following data
sample data:
partner cost
World T Plus 19000
Prix Mono 47500
Prix Mono 19000
Prix Mono 19000
World T Plus 19000
Prix Mono 19000
Prix Mono 38000
Prix Mono 19000
search string:
source="C:\tachiFlashDisc\test_firstAnswers.csv" |sort partner|table partner cost | dedup cost
result:
partner cost
Africa Telecom 9500
Africa Telecom 19000
Africa Telecom 47500
Africa Telecom 28500
B and C 0
B and C 4750
B and C 33250
B and C 14250
You could do something like this
| search_or_inputlookup_to_get_csv
| stats values(Products) as ProductList by TransactionID
| where isnotnull(mvfind(ProductList,"Product A"))