Hi List,

I wonder whether anybody has tried to draw the following idea with a sankey
diagram (but if you are aware of another way to draw the following, I am
open to suggestions):
I have several systems which are all linked so that part of the outputs of
each of the systems becomes part of the inputs of each of the others. So,
each system exchanges 2 flows with each of the other systems (an output
which becomes the other's input and vice-verse). This implies that some of
the flows will have to cross.

I tried many combinations to connect the flows to each other but I did not
manage to connect them all. In a 3 systems diagram, I managed at best to
connect 2 single flows: one from the 2nd diagram to the 1st and one from
the third to the second. The problem is that, no matter which order I gave
to the flows, I had never been able to connect both input and output
between any sub-diagram (maybe because matplotlib.sankey does not
automatically connect the flows when they have to cross?)

you can find the diagrams that I managed to draw here:
https://dl.dropbox.com/u/45960790/World/sankeys.7z

I would appreciate any help to connect all the remaining flows.

I think that this could be solved if it was allowed to explicitly state
more than one flow to connect but I have no clue whether it is possible to
implement:
Currently (as I have understood it), one can only specify a single "prior"
diagram and a single flow to be connected. So it would be nice to:
-- declare several flows to "connect" to the "prior" diagram; and also
-- declare several "prior" diagrams for which several flows could be
connected.

I guess this would be the easiest way to connect all flows from the
sub-diagrams.
Of course, this presupposes that crossing the flows is allowed. In case it
is not, can matplotlib.sankey be hacked to allow that?

Thanks a lot,

Aleix
------------------------------------------------------------------------------
Free Next-Gen Firewall Hardware Offer
Buy your Sophos next-gen firewall before the end March 2013 
and get the hardware for free! Learn more.
http://p.sf.net/sfu/sophos-d2d-feb
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Reply via email to