Is it possible to add multiple parameters to a Sankey diagram?
What do you mean parameters?
Parameters can be used for filters / in calculated fields.
Do you mean multiple sources / destinations? You can only have one source field and one destination field.
Max is correct. You can only use 2 dimensions, but if in your data you have values that are both a source and a destination you will wind up with outputs that look like this:
Hi @Jesse - how did you get a multiple-step sankey diagram?
I have data set of user events in an app that I want to visualize with sankey to show user journey in the app. I had use SQL to transform the data like following table
event_id, current screen_name, next screen_name,
then I use the sankey as follow, but it only show one source and one destination, but clearly in my dataset that there are values that are both source and a destination… why did it not show as multiple steps like yours?
Great question. I had been trying hard to achieve this but unable to do so.
Hi @Jesse your guidance will be great
Hi @luorobin and @Rajdipch, the data used to make that screenshot I attached looks like this. As long as you have values that are both source and destination (the destinations are sources of other rows, and so on and so on), you should get the effect of having multiple steps:
1 Like
Thanks @Jesse for the reply. I think my data is similar to yours, please see the screenshoot, where I highlight some of the values that are both source and destination.
@Jesse any thoughts on where I might be wrong?
I know that was just a sample set of rows, but appears that whenever NoteExportActivity is the source, NotepadActivity is the destination and vice versa. If you have instances where NoteExport is the source, Notepad is the destination, and then Notepad being a source with Login being the destination, then you would see Notepad showing up in the middle. Right now it is just a direct relationship between those 2 values even though it goes two ways.
1 Like
@Jesse I have data in the form of parent, child node similar to your dataset. But when I create sankey using the dataset I have it creates a one step sankey. Can you please explain if you did any custom calculations on quicksight?
Hi @rjnikita - no calculations needed, but in order to get multi-step sankeys you must have a given value as both a source and a destination. So you only need 3 columns, the source, destination, and value fields. If you see in the screenshot I posted further up, see ‘Electricity grid’ is a value both in source and destination columns, which will result in it being displayed in the ‘middle’.
Hi @Jesse , I have data regarding dog amount and tick amount. I need to show the association (source and targeted value) two values. in your screenshot. there is only one value shown… does it mean the value of source is presented by Frequency？(e.g. three dogA carry 4 ticks. five dogB carry 7ticks）
Hi @nimo - In general, Sankeys are really only for a single measurement and are meant to show the distribution of that measurement across N dimensions (in QS we support 2 dimensions). Sounds like you have 2 measures (num of dogs, num of ticks) and 1 dimension (dog type). For this a scatter plot would be best type of visualization in my opinion.