Modeling advise for processing mining

rks · ‎09-22-2021

Hi all,

in the latest blog post from the September 2021 update, there was a visual by PAFnow.

It collects the data from a event-like table and calculated lead times based on a key.

Let's say I want to track the movement a shipment. Each shipment has lots of scanpoints and corresponding scantypes:

ShipmentId	ScanType	Timestamp
124	LeaveWarehouse	2021-09-23 09:00:00
124	DepotEntry	2021-09-23 10:00:00
124	DepotSorted	2021-09-23 12:00:00
124	LoadOnTruck	2021-09-24 06:00:00
124	CustomerReceivedOK	2021-09-24 12:00:00
124	ReturnShipmentIssued	2021-09-25 16:00:00

So with a process mining tool I would be able to calculated the time between each node...

However, let's say I want to model the lead times by myself. Are there any best practises? Could I DAX the leadtimes (each measure is the process-step) with the table structure as provided above or should I pivot the table on loading based on the shipment?

I would highly appreciate to receive some information of blog posts or videos on similar use cases.

Best regards,

Konstantin

v-eqin-msft · ‎10-05-2021

Hi @rks ,

Could you tell me if my post helps you a little? If it is, kindly Accept it as the solution to make the thread closed. More people will benefit from it.

Best Regards,
Eyelyn Qin

rks · ‎10-06-2021

Hi @v-eqin-msft ,

I do like your links and ideas. And I also got the idea for asking about this topic by watching the recent blog entry. The solution provided there is also pafNow, which seems to be quite a bit of useful software. However, I was also interested in researching ways to achieve a smiliar, lest less generic, approach by using DAX only and not rely on external tools.