python 3 scatter plot gives "ValueError: Masked arrays must be 1-D" even though i am not using any masked array

Question

I am trying to plot a scatter plot using the .scatter method below. Here

ax.scatter(X[:,0], X[:,1], c = colors, marker = 'o', s=80, edgecolors = 'none')

with the input/args classes below:

X[:,0]] type: <class 'numpy.matrixlib.defmatrix.matrix'> X[:,1]] type: <class 'numpy.matrixlib.defmatrix.matrix'> colors type: <class 'list'>

however python is throwing a value error as seen here: error image

could you try to print out the shapes of the arrays? – J. P. Petersen May 28 '17 at 05:31 — J. P. Petersen, May 28 '17 at 05:31

score 25 · Accepted Answer · edited Jun 25 '17 at 09:23

25

Put the thing in brackets:

plt.scatter([X[:,0]],[X[:,1]])

edited Jun 25 '17 at 09:23

Dev-iL

23,742
7
57
99

answered Jun 23 '17 at 12:10

a good guy

266
3
2

Can you explain why this happens and why this fixes it? – May 30 '18 at 21:08
@EdgarAroutiounian see my answer for a possible explanation – Jeeter Jun 26 '18 at 15:02

score 9 · Answer 2 · answered Jun 26 '18 at 14:59

9

My experience with this is because your X is a numpy matrix.

Essentially, whenever you try and isolate a row from a matrix, it returns another matrix. Numpy seems to have a constraint that matrices must be 2 dimensional, so it can't tell that it's a 1-d array, and can't mask it (hence the Masked arrays must be 1-D error)

The solution for me was to simply "cast" X to a numpy.array by doing:

X = np.array(X)
ax.scatter(X[:,0], X[:,1], c = colors, marker = 'o', s=80, edgecolors = 'none')

answered Jun 26 '18 at 14:59

Jeeter

5,887
6
44
67

1

This fixed my issue when I encountered the same error message, thanks! – andrewwong97 Dec 05 '18 at 03:08
It also works when reduced to a list using numpy's tolist() function as in: X[:,0].tolist() – mikeTronix Feb 19 '20 at 20:13

python 3 scatter plot gives "ValueError: Masked arrays must be 1-D" even though i am not using any masked array

2 Answers2