Add Points To The Existing Matplotlib Scatter Plot

How to add points to the existing diagram? The straightforward solution is to plot a new scatter, adding new data. ax.scatter(data[:,0], data[:,1], cmap = cmap, c = color_data) ax

Solution 1:

It's unclear why creating a second scatter, as suggested by @b-fg, is not acceptable, but you could write a function like so:

def addPoint(scat, new_point, c='k'):
    old_off = scat.get_offsets()
    new_off = np.concatenate([old_off,np.array(new_point, ndmin=2)])
    old_c = scat.get_facecolors()
    new_c = np.concatenate([old_c, np.array(matplotlib.colors.to_rgba(c), ndmin=2)])

    scat.set_offsets(new_off)
    scat.set_facecolors(new_c)

    scat.axes.figure.canvas.draw_idle()

which allows you to add a new point to an existing PathCollection.

example:

fig, ax = plt.subplots()
scat = ax.scatter([0,1,2],[3,4,5],cmap=matplotlib.cm.spring, c=[0,2,1])
fig.canvas.draw()  # if running all the code in the same cell, this is required for it to work, not sure why
addPoint(scat, [3,6], 'c')
addPoint(scat, [3.1,6.1], 'pink')
addPoint(scat, [3.2,6.2], 'r')
addPoint(scat, [3.3,6.3], 'xkcd:teal')
ax.set_xlim(-1,4)
ax.set_ylim(2,7)

Note that the function that I'm proposing is very basic and would need to be made much smarter depending on the use case. It is important to realize that the facecolors array in a PathCollection does not necessarily have the same number of elements as the number of points, so funny things can happen with colors if you try to add several points as once, or if the original points are all the same colors, etc...

Solution 2:

To just add a new data with new colour, indeed calling again scatter will add the new points with the specified colour:

import numpy as np
import matplotlib.pyplotas plt

x = np.arange(10)
a = np.random.rand(10)
plt.scatter(x, a, c='blue')
b = np.random.rand(10)
plt.scatter(x, b, c='red')
plt.show()

Solution 3:

Assuming you already have a plot, you can create this function.

def AddPoint(plot, x, y, color):
    plot.scatter(x, y, c=color)
    plot.clf()
    plot.show()