Search code examples
pythonhistogramstacked-chart

how to plot two histograms with stacked bars, without stacking all the bars together?

I have 4 histograms, lets say A,B,C and D. I would like to plot histograms A and B together, with stacked bars, along with histograms C and D, also with stacked bars, but without stacking the four histograms together. So I would like two stacked histograms in a single histogram with side-by-side bars.

So far, I can plot either A-B-C-D with stacked bars; or A-B and C-D in different stacked histograms, but the bars of both histograms are not side by side. is the code I have:

plot=[A,B,C,D] #values from 0-10
ww=[wA,wB,wC,wD] #weights

All bars stacked:

plt.hist(plot,bins=10,weights=ww,label=['A','B','C','D'],histtype="barstacked")

A-B histogram + C-D histogram, but one histogram hides the other:

plt.hist(plot[0:2],bins=10,weights=ww[0:2],label=['loses','wins'],stacked=True)
plt.hist(plot[2:4],bins=10,weights=ww[2:4],label=['l','w'],stacked=True)

thanks in advance for your help!

Solution

  • you can use the patches list that hist() returns:

    import numpy as np
import matplotlib.pyplot as plt
A = np.arange(100)
B = np.arange(100)
C = np.arange(100)
D = np.arange(100)

wA = np.abs(np.random.normal(size=100))
wB = np.abs(np.random.normal(size=100))
wC = np.abs(np.random.normal(size=100))
wD = np.abs(np.random.normal(size=100))

plot=[A,B,C,D] #values from 0-10
ww=[wA,wB,wC,wD] #weights

n, bins, patches = plt.hist(plot[0:2],bins=10,weights=ww[0:2],label=['loses','wins'],stacked=True)
n, bins, patches2 = plt.hist(plot[2:4],bins=10,weights=ww[2:4],label=['l','w'],stacked=True)

for patch in patches:
    plt.setp(patch, 'width', 10)
for patch in patches2:
    plt.setp(patch, 'width', 5)    

plt.show()

    histogram

    update

    I found out there's a much better and cleaner way to do this:

    import numpy as np
import matplotlib.pyplot as plt

fig = plt.figure()
ax = fig.add_subplot(111)

# the data
A = np.arange(10)
B = np.arange(10)
C = np.arange(10)
D = np.arange(10)

wA = np.abs(np.random.normal(size=10))
wB = np.abs(np.random.normal(size=10))
wC = np.abs(np.random.normal(size=10))
wD = np.abs(np.random.normal(size=10))
## necessary variables
width = 0.5                      # the width of the bars
## the bars
rects1 = ax.bar(A - width/2, wA, width,
                color='blue')
rects2 = ax.bar(B- width/2, wB, width, bottom=wA,
                color='green')
rects3 = ax.bar(C + width/2, wC, width,
                color='red')
rects4 = ax.bar(D + width/2, wD, width, bottom=wC,
                color='yellow')
# axes and labels
ax.set_xlim(-width,len(A)+width)

plt.show()

    the results it: better histogram

    for more details look at this link.