Sankey diagram - data structure question

[RichardP]

Hi,

I'm attempting to see if I can use the Sankey visual to illustrate the main user journeys through a website.  However I'm not sure what format I need the data in to make this work.

The current format of the data is:

USER ID	DATETIME	WEBSITE SECTION	POSITION IN JOUNREY
111	20/10/2017 09:18:21	Home	1
111	20/10/2017 09:18:30	Search Results	2
111	20/10/2017 09:19:00	Article Page	3
111	20/10/2017 09:19:58	Article Page	4
111	01/11/2017 17:19:58	Article Page	1
111	01/11/2017 17:22:58	Home Page	2
111	01/11/2017 17:23:25	Article Page	3
222	07/10/2017 12:14:25	Home Page	1
222	07/10/2017 12:15:10	Search Results	2
333	24/10/2017 15:15:10	Home Page	1
333	24/10/2017 15:15:25	Article Page	2
333	24/10/2017 15:16:01	Home Page	3

The "Position in Journey" column identifies that it was the first, second, etc page to be visited during a visit.

A given user may make multiple visits to the website.

A successful outcome for me would be to show the pages in 'Position 1', with the flows on to the pages in 'Position 2', etc.

If there are better ways than a Sankey visualisation to achieve this, I'm happy to learn! 🙂

[v-sihou-msft · Accepted solution]

@RichardP

To use Sankey chart, you dataset should have columns like Source, Target, Weight column. It can be like: 

It means on row level, you need to have current Position and next Postion info, with some weight. So in your scenario, you need to put Positions into Source and Target. It can be like: 

Source                 Target 

Home                  Search Results

Search Results        Article Page

Article Page         Article Page

However, this visual currently can't analyze on some other data field level.

As I notice that you also have UserID and DateTime columns in your table, you may create one slicer for User, one slicer for DateTime. 

Currently I haven't seen any other published custom visual better on showing flow diagram. 

Regards,