使用 matplotlib 绘制包含数千个点及其标签（组）的数据

Question

I have an array of 500000 samples ie, the data's shape is (500000, 3) where the first two columns represent x-coordinate and y- coordinate, and the third column is Label values to which the datapoint @ (X,Y) belongs.

for example:- data= [ [20,10, 12.3320], [22, 13, 230.221],.....[..] ]

I tried the below method. But this is too time consuming and poorly interpreted.

import matplotlib.pyplot as plt

colors = 10*['r.','g.','b.','c.','k.','y.','m.']

for i in range(len(labels)):
    plt.scatter(data[i][0], data[i][1], colors[labels[i]],marker='.')

plt.show()

Is there any other method like imshow() or other which is suitable for the above code which leads to good interpretation?

Answer 1

The scatter function in matlplotlib is quiet slow, I would recommend to use vispy that use the GPU to plot a large number of points :

Works with vispy 0.4.0 that you can install with pip or conda :

pip install vispy

Here is the code (plotted in less than 2sec on my computer):

import numpy as np
from vispy import scene, visuals, app
import matplotlib.pyplot as plt

data = np.random.random((500000,3))

canvas = scene.SceneCanvas(keys='interactive', show=True)
view = canvas.central_widget.add_view()
# Create the scatter plot
scatter = scene.visuals.Markers()
scatter.set_data(data[:,:2], face_color=plt.cm.jet(data[:,2]))
view.add(scatter)
view.camera = scene.PanZoomCamera(aspect=1)
view.camera.set_range()
app.run()

there is a nice documentation for vispy and you can customize your plot in the set_data function with arguments like face_color, edge_color, size, edge_width, symbol ...

Good luck with your data visualization ;)