I have a data set of transactions in a messaging pool. I am unable to share the workbook due to protected information, but here is a very simple sample mockup:
Sender | Recipient | Timestamp | Thread ID
Anna | Benjamin | 2/6/25 16:36 | 12345
Benjamin | Carol | 2/6/25 17:30 | 12345
Carol | Benjamin | 2/6/25 18:20 | 12345
Benjamin | Anna | 2/7/25 10:15 | 12345
Each "conversation" (Thread ID) can have any number of overall participants in that thread, but I want to be able to get statistics (average, maximum, standard deviation) of response times per person from when that person receives a message (as Recipient) to when they take action to send the next related message (as Sender)
Thanks!
Hi @Seth Noble
Some comments/questions:
1) What time value are you looking to return back to then aggregate upon? In 'seconds', 'minutes', 'hours'? Also, if the time value is greater than 24 hours then we're getting into a completely different situation to return a proper result. Is that a possibility in your data?
2) A broader set of example data might be needed here with different Thread IDs.
3) The request will involve use of LOOKUP() function which is a table calculation. While useful and powerful, table calc's carry limitations (1) table calculations provide a displayed value, not an actual value to work with going forward, such as incorporating the result in another calculation (2) attempts to remove or modify the structure of the table could cause the calculation to error or the table to collapse. Such as when creating a bar/line chart or aggregating/summarizing thereafter.
Those are the caveats, the watch-outs, when using table calculations. And, we won't be able to provide a grand total easily with this function (see below screenshot), if at all, which is what you ultimately need to work with:
So, this request would be best performed in advance of Tableau Desktop in an ETL tool such as Tableau Prep to provide you with a real value to work with in Desktop thereafter.
Best, Don Wise -
Please don’t forget to upvote and/or Select as Best by clicking the hyperlink below in the response that answered your question.
