Group scatter points by a certain condition and make a legend from it

Question

I have a number of data I want to plot with every combination in 3D. So Data looks like that.

I created a matrix:

a = array([[1,2,2],[3,4,3],[4,2,2],[2,7,4],[5,8,1],[7,1,7]])

Im gonna skip the loop for this example.

k = 1 ; j = 0; z = 2
fig = figure()
ax = fig.add_subplot(111, projection='3d')
ax.scatter(a[:,j], a[:,z], a[:,k])

Next step is that I wanna color plots by the column A (if they belong to Random1 or Random2) and make a legend based on this condition. And it doesnt work.

My attempt was to make a list of column A

a = [Random1, Random2]

and put it in my figure ploting.

k = 1 ; j = 0; z = 2
fig = figure()
ax = fig.add_subplot(111, projection='3d')
ax.scatter(a[:,j], a[:,z], a[:,k], label = a)
ax.legend()

basically I want to have something like this scatter plot with legend colored by group without multiple calls to plt.scatter

but with matplotlib (its a necessity) and in 3d. Hope its clear what I want to do.

Edit: with this addition, it does color them as I want, but Still have no idea who to make a legend based on the color.

ax.scatter(a[:,j], a[:,z], a[:,k], c = len, cmap = 'brg'))

where len is length of my data.

Can I use something like pandas, to add labels to corresponding data, but read data from matrix?

What do you mean as points? Isn't that scatter about to plot them as points not connect them? This is just an example. I have around 10000 points and around 30 groups of random in my data. — Noob Programmer, Nov 16 '19 at 12:44
or something like this, but again with matplot lib and in 3d. - https://stackoverflow.com/questions/44595288/matplotlib-adding-legend-based-on-existing-color-series — Noob Programmer, Nov 16 '19 at 12:59
you want to plot each line or all possible combinations? from the loop you use I believe you need all combinations but then the labels do not make sense. Line 1 has label Random1 but if you mix the points what are the labels? — seralouk, Nov 16 '19 at 14:45
I want to plot possible combination B to C, B to D, B to E(data is huge) and label them by column A. So first 3 combinations, so in our case have x1,y1,z1 = (1,2,2), x2,y2,z2 = (3,4,3), x2,y2,z2 = (4,2,2), are lets say yellow color and have a label Random 1. — Noob Programmer, Nov 16 '19 at 14:51

seralouk · Accepted Answer · 2019-11-16T15:33:24.317

I am updating my answer. So based on what you said you want to plot all possible combinations of the points.

This should work fine

from mpl_toolkits.mplot3d import Axes3D
import matplotlib.pyplot as plt
from numpy import array
import matplotlib.patches as mpatches

a = array([[1,2,2],[3,4,3],[4,2,2],[2,7,4],[5,8,1],[7,1,7]])

colors = ['r', 'r','r','g','g','g']

fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')

for i in range(3):
    for j in range(3):
        for k in range(3):
            ax.scatter(a[:,i], a[:,j], a[:,k], c=colors)

ax.set_xlabel('X Label')
ax.set_ylabel('Y Label')
ax.set_zlabel('Z Label')

R1 = mpatches.Patch(color='red', label='Random 1')
R2 = mpatches.Patch(color='green', label='Random 2')
plt.legend(handles=[R1,R2])
plt.show()

Group scatter points by a certain condition and make a legend from it

1 Answers1