Change colour of colorbar in Python Matplotlib

Question

I have a code that gives me a scatter plot of predicted vs actual values as a function of concentration. The data is pulled from an excel csv spreadsheet.

This is the code:

import matplotlib.pyplot as plt
from numpy import loadtxt


dataset = loadtxt("ColorPlot.csv", delimiter=',')
x = dataset[:,0]
y = dataset[:,1]
z = dataset[:,2]

scaled_z = (z - z.min()) / z.ptp()

colors = plt.cm.viridis(scaled_z)
sc=plt.scatter(x, y, c=colors)
plt.clim(0, 100)
plt.colorbar()
plt.xlabel("Actual")
plt.ylabel("Predicted")

plt.show()

And with this I get a nice graph:

However if I change the color to something like

colors = plt.cm.plasma(scaled_z)

I get the graph below but the colorbar remains unchanged.

I've tried lots of different things like cmap or edgecolors but I don't know how to change it. And I want to keep the code as simple as it currently is because I want to readily change the third variable of z based on my excel spreadsheet data.

Is there also a way for the scale of the colorbar to pick up what the scale is from the excel spreadsheet without me manually specifying 0-100?

Answer 1

To get the right color bar, use the following code:

colormap = plt.cm.get_cmap('plasma') # 'plasma' or 'viridis'
colors = colormap(scaled_z)
sc = plt.scatter(x, y, c=colors)
sm = plt.cm.ScalarMappable(cmap=colormap)
sm.set_clim(vmin=0, vmax=100)
plt.colorbar(sm)
plt.xlabel("Actual")
plt.ylabel("Predicted")
plt.show()

For my random generated data I got the following plot:

Now replace 'plasma' with 'viridis' and check the other variant.

Answer 2

You should not scale your data, unless you want the colorbar to be incorrect. Once you have the PathCollection from the scatter call, you can call set_cmap and set_clim on that and the colorbar should track. (you could also explicitly associate the colorbar with the PathCollection to avoid ambiguity)

import matplotlib.pyplot as plt
import numpy as np


x = np.random.randn(100)
y = np.random.randn(100)
z = np.random.randn(100)

sc=plt.scatter(x, y, c=z, cmap='viridis')
plt.clim(0, 100)
plt.colorbar(sc)
plt.xlabel("Actual")
plt.ylabel("Predicted")
sc.set_cmap('plasma')
sc.set_clim(-1, 1)

plt.show()