Hello everyone, I have a question regarding the color bar legend for graphs involving continuous color scales.
I was wondering if there is a way for me to reverse the order of the color bars, such that the lowest numbers are on the top and the highest numbers are on the bottom.
I have tried using:
fig.update_coloraxes(reversescale=True)
However, it only reversed the color scale and not the numerical scale.
I am more so trying to accomplish something like this:
Hi @Conor ,
As far as I know, that feature is not available in built-in right now.
But it does not mean impossible.
The possible solution that came up into my mind is by creating 2 same plot objects , Scatter as example.
After that, by using those 2 plot objects you can :
labelalias attribute.reversescale=True and hide the colorbar of second plot using showscale=False .Here the code of those scenario
import plotly.graph_objects as go
import numpy as np
y_data = [10,20,30,40,50,60,70,80,90,100]
x_data = [10,20,30,40,50,60,70,80,90,100]
tickvals = [0, 25, 50, 75, 100]
reverse_labelalias = dict([(k,str(v)) for k,v in zip(tickvals,sorted(tickvals, reverse=True))])
fig = go.Figure(data=[
# First plot to show the colorbar only
go.Scatter(
y = y_data,
x = x_data,
mode='markers',
name='',
marker=dict(
size=16,
color=y_data,
showscale=True,
colorbar=dict(
title="Rank",
tickmode="array",
tickvals= tickvals,
labelalias= reverse_labelalias,
)
)
),
# Second plot to show the trace only
go.Scatter(
y = y_data,
x = x_data,
mode='markers',
name='Rank',
marker=dict(
size=16,
color=y_data,
reversescale=True,
showscale=False,
colorbar=dict(
title="Rank",
tickmode="array",
)
)
)
]
)
fig.update_layout(showlegend=False, title="Reverse Range Scale")
fig.show()
Hope this help you.
I am trying this idea out right now. My interface is using px, so I have been trying to re-work this solution with plotly express in mind. Right now I am trying to overlay the coloraxis/colorbar from another graph, however, it seems to not be working as expected.
As of now the problem is that the colorbar from legendTrace isn’t showing. Is there something that I am doing wrong that prevents this from working?
My code is as follows:
fig = px.scatter(df_melted, x="value", facet_col_spacing=0.04, y="y", facet_col="feature",color='rank',color_continuous_scale='Plasma_r', facet_col_wrap=2,custom_data=['id'], hover_data={'id': True}, category_orders={
"feature": [feature + "_v" for feature in features]})
fig.update_coloraxes(showscale=False)
fig.update_yaxes(matches=None)
trendFig = px.scatter(df_melted, x="value", facet_col_spacing=0.04, y="y", facet_col="feature", facet_col_wrap=2, trendline="ols", category_orders={
"feature": [feature + "_v" for feature in features]})
legendTrace = px.scatter(df_melted, x="value", facet_col_spacing=0.04, y="y", facet_col="feature",color='rank',color_continuous_scale='Plasma', facet_col_wrap=2,custom_data=['id'], hover_data={'id': True}, category_orders={
"feature": [feature + "_v" for feature in features]})
trendFig.update_traces(visible=False, selector=dict(mode="markers"))
legendTrace.update_traces(visible=False, selector=dict(mode="markers"))
legendTrace.update_coloraxes(showscale=True)
fig.add_traces(trendFig.data)
fig.add_traces(legendTrace.data)
I apologize if my code is a bit messy, I am fairly new to dash/plotly as a whole.
Hi @Conor ,
Try to change last five lines into lines below:
trendFig.update_traces(visible=False, selector=dict(mode="markers"))
# comment a line below
# legendTrace.update_traces(visible=False selector=dict(mode="markers"))
legendTrace.update_coloraxes(showscale=True)
# switch position 2 last lines
fig.add_traces(legendTrace.data)
fig.add_traces(trendFig.data)
Unfortunately it is still not working; would you know of an alternative way of going about this?
Hi @Conor ,
Here the code with Plotly Express version.
import plotly.express as px
import pandas as pd
df = pd.DataFrame({'x_data':[10,20,30,40,50,60,70,80,90,100],'y_data':[10,20,30,40,50,60,70,80,90,100]})
tickvals = [0, 25, 50, 75, 100]
reverse_labelalias = dict([(k,str(v)) for k,v in zip(tickvals,sorted(tickvals, reverse=True))])
# First plot to show the colorbar only
fig = px.scatter(df, x="x_data", y="y_data")
fig.update_traces(marker=dict(
size=16,
color=df["y_data"],
showscale=True,
colorbar=dict(
title="Rank",
tickmode="array",
tickvals= tickvals,
labelalias= reverse_labelalias,
)
))
# Second plot to show the trace only
fig1 = px.scatter(df, x="x_data", y="y_data")
fig1.update_traces(name='Rank',
marker=dict(
size=16,
color=df["y_data"],
reversescale=True,
showscale=False,
colorbar=dict(
title="Rank",
tickmode="array",
)
)
)
fig.add_traces(fig1.data)
fig.update_layout(showlegend=False, title="Reverse Range Scale")
fig.show()
This seems to be working on a colab notebook! However, when I implemented the concept into my code, I get this weird error:
ValueError: Invalid property specified for object of type plotly.graph_objs.scatter.marker.ColorBar: ‘labelalias’
legendFig.update_traces(marker=dict(
color=df_melted["rank"],
showscale=True,
colorbar=dict(
title="Rank",
tickmode="array",
labelalias = reverse_labelalias,
tickvals = tickVal,
)
))
I am not sure what could be causing this; since the colab notebook is not having this issue.
Hi @Conor ,
Please check the version of your Plotly.
it is possible that the problem is version gap issue.