CODEWITHSUNDEEP
pythonmatplotlibdata-sciencemnist
IsaacLevon
IsaacLevon

Reputation: 2570

plotting MNIST samples

I'm trying to plot 10 samples from the MNIST dataset. One of each digit. Here's the code:

import sklearn
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
from sklearn import datasets

mnist = datasets.fetch_mldata('MNIST original')
y = mnist.target
X = mnist.data

for i in range(10):
    im_idx = np.argwhere(y == i)[0]
    print(im_idx)
    plottable_image = np.reshape(X[im_idx], (28, 28))
    plt.imshow(plottable_image, cmap='gray_r')
    plt.subplot(2, 5, i + 1)

plt.plot()

For some reason, the zero digit is being skipped in the plot.

Why?

Upvotes: 2

Views: 5807

Answers (2)

Sheldore
Sheldore

Reputation: 39042

Ok, I got it. The problem was that you were defining the subplot after plotting imshow. So your first subplot was overwritten by the second one. To make your code work, just swap the order of your two commands as following. Also, I don't see why you are using plt.plot() at the end.

plt.subplot(2, 5, i + 1) # <-- You have put this command after imshow 
plt.imshow(plottable_image, cmap='gray_r')

Here is another alternate for your knowledge:

fig = plt.figure()

for i in range(10):
    im_idx = np.argwhere(y == i)[0]
    plottable_image = np.reshape(X[im_idx], (28, 28))
    ax = fig.add_subplot(2, 5, i+1)
    ax.imshow(plottable_image, cmap='gray_r')

You can also further shorten Scott's code (posted below) by using the following:

fig, ax = plt.subplots(2,5)
for i, ax in enumerate(ax.flatten()):
    im_idx = np.argwhere(y == i)[0]
    plottable_image = np.reshape(X[im_idx], (28, 28))
    ax.imshow(plottable_image, cmap='gray_r')

enter image description here

Upvotes: 4

Scott Boston
Scott Boston

Reputation: 153460

Try this:

import sklearn
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
from sklearn import datasets

mnist = datasets.fetch_mldata('MNIST original')
y = mnist.target
X = mnist.data

fig, ax = plt.subplots(2,5)
ax = ax.flatten()
for i in range(10):
    im_idx = np.argwhere(y == i)[0]
    print(im_idx)
    plottable_image = np.reshape(X[im_idx], (28, 28))
    ax[i].imshow(plottable_image, cmap='gray_r')

Output:

enter image description here

Upvotes: 4

Related Questions