Conditional marker matplotlib

Question

I plot the result of a Gaussian Mixture when I have 3 type of points that are determinate by a prediction. I have a different color for each cluster predicted and now I would like to have a different marker instead of color.

colors=['pink' if i==0 else 'skyblue' if i==1 else 'lightgreen' for i in resultGM]
markers=['c' if i==0 else 'o' if i==1 else 'D' for i in resultGM]
ax=plt.gca()
ax.scatter(datas[:, 0], datas[:, 1], alpha=0.8, c=colors, marker=markers)   
plt.title("Calling " + name_snp)
plt.xlabel('LogRatio')
plt.ylabel('Strength')
plt.show()

It works perfectly for colors like this:

But I can't do the same thing with different markers, it doesn't recognize a list of markers.

How can I do to have a different marker for each cluster (0,1,2) like I have with colors?

You have to loop through them. See here http://stackoverflow.com/questions/41078331/matplotlib-read-marker-direction-from-a-file/41078504#41078504 — p-robot, Dec 12 '16 at 11:35

score 3 · Accepted Answer · answered Dec 12 '16 at 11:36

3

Change the line with plt.scatter in it to this instead:

for x, y, c, m in zip(datas[:,0], datas[:,1], colors, markers)
    ax.scatter(x, y, alpha=0.8, c=c,marker=m)

answered Dec 12 '16 at 11:36

p-robot

4,652
2
29
38

Please don't call `scatter` for each data point separately. This will make your plot unmanageable. – Mad Physicist Dec 22 '19 at 10:16
It's not called for each data point. It's called for each colour in the dataset. – p-robot Jan 15 '20 at 12:06

score 1 · Answer 2 · answered Dec 22 '19 at 10:25

Assuming that resultGM is an array of integer labels. If it's not, turn it into one. You should have exactly one call to scatter for each cluster, not one call per data point as the other answer suggests:

colors = ['pink', 'skyblue', 'lightgreen', ...]
markers = ['c', 'o', 'D', ...]

for i in range(resultGM.max()):
    mask = (resultGM == i)
    ax.scatter(data[mask, 0], data[mask, 1], alpha=0.8, c=colors[i], marker=markers[i])

Conditional marker matplotlib

2 Answers2

Linked