3d plot of list of (hist, bin_edges) where histogram bar chart or lines are in the z-y plane

Question

EDIT - reworked question

I need to print a 3D histogram of fitness data for 50 generations of computer programmes. This data is in calculated and stored in a logbook using the DEAP framework. The form of the plot needs to be with the fitness frequency on the z axis, generation on the x axis and bin_edges on the y axis. So the lines for the histogram are in the z-y plane, for each generation on the x-axis.

The frequency data for each generation is contained in a numpy array of shape (#generations, #bin_edges), gained by running np.histogram() on each generation.

histograms = [el[0] for el in logbook.chapters["fitness"].select("hist")]

histograms.shape
(51, 10)  # (num gen, num bins)

print (histograms)  # excerpt only
[[ 826.  145.   26.    2.    1.    0.    0.    0.    0.    0.]
 [ 389.  446.  145.   16.    4.    0.    0.    0.    0.    0.]
 [ 227.  320.  368.   73.   12.    0.    0.    0.    0.    0.]
 [ 199.  128.  369.  261.   43.    0.    0.    0.    0.    0.]
 [ 219.   92.  158.  393.  137.    1.    0.    0.    0.    0.]
 [ 252.   90.   91.  237.  323.    6.    1.    0.    0.    0.]
 [ 235.   89.   69.   96.  470.   36.    5.    0.    0.    0.]
 [ 242.   78.   61.   51.  438.  114.   16.    0.    0.    0.]
 [ 235.   82.   52.   52.  243.  279.   57.    0.    0.    0.]]

bin_edges
array([  0.,   9.,  18.,  27.,  36.,  45.,  54.,  63.,  72.,  81.,  90.])

gen
[0, 1, 2, 3, 4, 5, 6, 7, 8, ...]

I've made several attempts, but can't seem to get the histogram data into the right format, or possibly shape for matplotlib axes.bar.

ATTEMPT 2: busy reworking

import matplotlib.pyplot as plt
import matplotlib as mpl
import numpy as np
xedges = list(gen)
yedges = list(bin_edges)
H = histograms.T

fig=plt.figure()
# error on this line: TypeError: list indices must be integers or slices, not list
ax = fig.add_subplot(133, title='NonUniformImage: interpolated', aspect='equal', \

xlim=xedges[[0, -1]], ylim=yedges[[0, -1]])
im = mpl.image.NonUniformImage(ax, interpolation='bilinear')
xcenters = (xedges[:-1] + xedges[1:]) / 2
ycenters = (yedges[:-1] + yedges[1:]) / 2
im.set_data(xcenters, ycenters, H)
ax.images.append(im)
plt.show()

ATTEMPT 1:

fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
# I have a list of len(gen) histograms
# with an array of freq count for each bin

xs =  list(gen)
ys = list(bin_edges)
zs = hist.T #ndarray


# error occurs here as 
# ValueError: shape mismatch: objects cannot be broadcast to a single shape    
ax.bar(xs, ys, zs)
plt.show()

I'm trying to work through the mplot3 hist3d demo https://matplotlib.org/examples/mplot3d/hist3d_demo.html — Blue Shrapnel, Dec 30 '17 at 22:48
A similar question https://stackoverflow.com/questions/47899253/how-to-plot-3d-histogram-of-an-image-in-opencv/47899963#47899963 — Kinght 金, Dec 31 '17 at 05:33
Thanks @Silencer, I've tried to run that, but there's an error importing cv2, `Symbol not found: _clock_gettime`. I think the problem is in the structure of the histogram data, as it's in the z-y plane, for each bin in x. All the examples I've seen have the lines in the z-x plane. — Blue Shrapnel, Dec 31 '17 at 09:41
@BlueShrapnel The OpenCV is needed in the my post link, but is not necessary for yours. And you didn't give your histogram data format and content. It's hard to tell why. — Kinght 金, Dec 31 '17 at 09:49
histogram sample data is in the beginning of the post... histograms — Blue Shrapnel, Dec 31 '17 at 09:58
@Silencer, I wanted to see the format of the output from calcHist, to see if I could find what was wrong with my histogram data. — Blue Shrapnel, Dec 31 '17 at 09:59
@BlueShrapnel Look this carefully, is it right as your attention? https://i.stack.imgur.com/EdpOA.png — Kinght 金, Dec 31 '17 at 10:50
@Silencer, thank you. Very close, need to swap bins and idx. — Blue Shrapnel, Dec 31 '17 at 11:15
@BlueShrapnel Is this ok? https://i.stack.imgur.com/hikNN.png — Kinght 金, Dec 31 '17 at 11:26

score 2 · Accepted Answer · answered Dec 31 '17 at 11:45

I use mplot3d and bar to plot 3d-hist as follow:

#!/usr/bin/python3
# 2017.12.31 18:46:42 CST
# 2017.12.31 19:23:51 CST
import numpy as np
from mpl_toolkits.mplot3d import Axes3D
import matplotlib.pyplot as plt
import numpy as np

## the hist data
data = np.array([
        np.array([826, 145,  26,   2,   1,   0,   0,   0,   0,   0]),
        np.array([389, 446, 145,  16,   4,   0,   0,   0,   0,   0]),
        np.array([227, 320, 368,  73,  12,   0,   0,   0,   0,   0]),
        np.array([199, 128, 369, 261,  43,   0,   0,   0,   0,   0]),
        np.array([219,  92, 158, 393, 137,   1,   0,   0,   0,   0]),
        np.array([252,  90,  91, 237, 323,   6,   1,   0,   0,   0]),
        np.array([235,  89,  69,  96, 470,  36,   5,   0,   0,   0]),
        np.array([242,  78,  61,  51, 438, 114,  16,   0,   0,   0]),
        np.array([235,  82,  52,  52, 243, 279,  57,   0,   0,   0])
        ])

## other data 
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
colors = ["r","g","b"]*10

## Draw 3D hist 
ncnt, nbins = data.shape[:2]
xs = np.arange(nbins)
for i in range(ncnt):
    ys = data[i]
    cs = [colors[i]] * nbins
    ax.bar(xs, ys.ravel(), zs=i, zdir='x', color=cs, alpha=0.8)

ax.set_xlabel('idx')
ax.set_ylabel('bins')
ax.set_zlabel('nums')
plt.show()

Thank you! It's definitely a black art which I don't yet understand. — Blue Shrapnel, Dec 31 '17 at 12:07
How can I change the labels of the bin axis? I've tried `ax.set_xticklabels(bin_labels[1:])` where bin_labels is `['0', '9', '18', '27', '36', '45', '54', '63', '72', '81', '90']`. But there is no change. — Blue Shrapnel, Dec 31 '17 at 12:47
Modify this line: `xs = xxx` as you want, take care of the dimension. — Kinght 金, Dec 31 '17 at 13:02

3d plot of list of (hist, bin_edges) where histogram bar chart or lines are in the z-y plane

1 Answers1

Linked