Imshow heatmap with array of arrays - problem with final plot/image-CodePudding

I'd like to plot imshow heatmap but the final image looks very small and disproportional. My data represent a distance measured between atoms (cca 10 atoms of one molecule to cca 30 atoms of other molecule) - the result is array of arrays. I prepared similar output for illustration, however my original dataset is bigger:

import numpy as np
array1 = np.random.randint(20, size=30)
array2 = np.random.randint(20, size=30)
array3 = np.random.randint(20, size=30)
array4 = np.random.randint(20, size=30)
array5 = np.random.randint(20, size=30)
array6 = np.random.randint(20, size=30)
array7 = np.random.randint(20, size=30)
array8 = np.random.randint(20, size=30)
array9 = np.random.randint(20, size=30)
arrayOfArrays = np.array([array1, array2, array3, array4, array5, array6, array7, array8, array9])

Then I wanted to make heatmap to see distances between the atoms, so this is my code:

import matplotlib.pyplot as plt

fig, ax = plt.subplots()
im = ax.imshow(arrayOfArrays, origin='upper')

#this is here because I use this approach to define xticks and yticks in my original plot  - here I modified the code with "len(range(0,30))" but in my original plot there is number of atoms for which I measured the distances (something like n_atoms=len(dataset1))
n_1=len(range(0,30))
n_2=len(range(0,9))
tick_interval = 1
ax.set_yticks(np.arange(n_2)[::tick_interval])
ax.set_xticks(np.arange(n_1)[::tick_interval])

# colorbar
cbar = fig.colorbar(im)

The final image created with this particular plot looks already small but the image with my original data is even smaller and I can't see the colors inside at all.

I'd really appreciate any suggestions where the problem is or which part of the code should be edited? I tried to edit size of picture, add "interpolation", "ascpect"...

CodePudding user response：

First of all, there is no need to create your arrays like that. Instead, you can use list comprehension to get everything in one line:

arrayOfArrays = np.array([np.random.randint(20, size=30) for _ in range(9)])

I know that you are probably using a dataset, however, this is useful to know for creating the data randomly :3

Also there are some other redundancies in your code, you can use

np.arange(30)

Instead of using

n_1 = len(range(0,30))
np.arange(n_1)

to create a list of integers from 0 to 30

Matplotlib figure resizing: You can adjust the size of the figure, everything that is plotted with

fig.size_in_inches(10, 10)

When I insert this line in your code and run it in google colab, the output is as follows:

If you want to adjust the size of the color bar use the shrink argument in the fig.colorbar, a value of 0.3 outputs:

CodePudding user response：

I like the answer from TomiOck, but I believe that JohanC is right is you have a dataset for which you have many more columns than rows. This can be fixed by adding aspect='auto' to the imshow command.

Here is my attempt:

import numpy as np
nrows = 9
ncols = 30
minv = 0 # inclusive
maxv = 20 # exclusive
arrayofArrays = np.random.randint(low=minv, high=maxv, size=(nrows,ncols))

You can generate the random data in one numpy line.

import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize=(14,8)) # width and height in inches
im = ax.imshow(arrayOfArrays, origin='upper', aspect='auto', interpolation='None')

n_1 = np.arange(ncols)
n_2 = np.arange(nrows)
tick_interval = 1
ax.set_yticks(n_2[::tick_interval])
ax.set_xticks(n_1[::tick_interval])

# colourbar
cbar = fig.colorbar(im)

The aspect='auto' stretches your image such that it fill the figure. I also recommend using interpolation='None' to be absolutely sure that matplotlib does not change the colours much (