Plotly Sankeys fill all the available space.
If some "branches" have 3 levels (Node 0 -> Node 1 -> Node 3) and others only two (Node 0 -> Node 2), the final nodes (Node 2 and Node 3) are vertically aligned to the right.
I would like to be able to align vertically Node 1 (not Node 3!) and Node 2, as in the second figure in attachment.
It is quite difficult for me to align all nodes programmatically, so I would like to avoid this possibility.
Even in the former undesired case something goes wrong (code for Jupyter Lab):
import plotly.graph_objects as go
sources = [ 0, 0, 1, 1]
targets = [ 1, 2, 3, 4]
values = [59, 30, 29, 30]
labels = ['Node 0', 'Node 1', 'Node 2', 'Node 3', 'Node 4']
link = dict(source=sources, target=targets, value=values)
node = dict(label=labels)
data = go.Sankey(link=link, node=node)
fig = go.Figure(data)
fig.show(renderer="svg", width=1000, height=500)
produces:
It looks like there is a bug when placing Node 1 that doesn't take in account the thickness of Node0->2 flow.
Is it possible to tell that I want Node 2 on the same vertical line of Node 1, and that Node 2 has no children? A solution would be the manual placement, alas I cannot do it because my Sankey is highly dynamic.
As a workaround I created a dummy node after Node 2, and made it invisible:
import plotly.graph_objects as go
sources = [ 0, 0, 1, 1, 2]
targets = [ 1, 2, 3, 4, 5]
values = [59, 30, 29, 30, 30]
labels = ['Node 0', 'Node 1', 'Node 2', 'Node 3', 'Node 4']
color_links = ["lightgray", "lightgray", "lightgray", "lightgray", "white"]
color_nodes = ["red", "blue", "green", "navy", "lightgreen", "white"]
link = dict(source=sources, target=targets, value=values, color=color_links)
node = dict(label=labels, color=color_nodes, line = dict(width = 0))
data = go.Sankey(link=link, node=node)
fig = go.Figure(data)
fig.show(renderer="svg", width=1000, height=500)
this produces:
I would expect a better and easier way to do this. Is there any pointer to a working demo?
from Plotly Sankey: how to avoid autoplacing all nodes to the right?
No comments:
Post a Comment