Well, I don't think your data structure is suitable for path analysis. A sankey diagram is a bit like a network diagram as it shows lines (paths) between nodes (events) appearing over time. So for instance an event (e.g. a product purchase, a registration, a cancellation, etc.) which occurs over and over again during a specific time period. A sankey diagram would show you the number of people going thru such an event - so you can measure effectiveness or general volume.
In your case - you have a transaction data base with a distinct event per transaction. I'm not sure what output you would expect given your sample data - but given your 4x senders (transaction identifiers) and receivers (events) - it would render like this:
Given a receiver doesn't appear more than once in your data across different senders - the sankey never connects paths here. Hence you have such horizontal lines here - one for each sender.
You may also consider using a network visualization instead. Given your 4x senders and amount sent to an address - it may render like this:
It doesn't do too well with dates/times but may still be useful to see overall relationship between senders/receivers.
Sankey diagrams are mostly used to visualize paths (hence you find this under path analysis in VA) - so it may not be as useful for visualizing a distinct list of unique transaction details.
Hope this helps. Regards, Falko
... View more
