Matplotlib/seaborn histogram using different colors for grouped bins

Question

I have this code, using a pandas df:

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import os


path_to = 'Data\\2017-04\\MonthlyData\q1analysis\Energy Usage'  # where to save

df = pd.read_csv('April2017NEW.csv', index_col =1)


df1 = df.loc['Output Energy, (Wh/h)']  # choose index value and Average
df1['Average'] = df1.mean(axis=1)
print df1
print df1['Average'].describe()

def hist():
    p = sns.distplot(df1['Average'],kde=False, bins=25).set(xlim=(0, 100));
    plt.xlabel('Watt hours')
    plt.ylabel('Households')

    return plt.show()

which returns:

I would like to use three different colors (low, medium, high) to represent higher values on the x =axis with a legend, like this:

EDIT1:

I found this example: here, so I am trying to use this.

I've come up with this: Which is almost there. How does one split the range into 3, with 3 different colors?

score 6 · Accepted Answer · answered May 09 '17 at 14:03

6

Solution:

N, bins, patches = plt.hist(df1['Average'], 30)

cmap = plt.get_cmap('jet')
low = cmap(0.5)
medium =cmap(0.2)
high = cmap(0.7)


for i in range(0,3):
    patches[i].set_facecolor(low)
for i in range(4,13):
    patches[i].set_facecolor(medium)
for i in range(14,30):
    patches[i].set_facecolor(high)

plt.xlabel("Watt Hours", fontsize=16)  
plt.ylabel("Households", fontsize=16)
plt.xticks(fontsize=14)  
plt.yticks(fontsize=14)
ax = plt.subplot(111)  
ax.spines["top"].set_visible(False)  
ax.spines["right"].set_visible(False)

plt.show()

output:

answered May 09 '17 at 14:03

warrenfitzhenry

2,209
8
34
56

1

There is an error in your code (and also in the picture, look closely!). When you call range(0,3) you get the numbers 0,1,2. No 3! So when you set the facecolor of these patches to `low`, you miss out the patch with index 3. So either use range(0,4) for low and range(4,13) for medium or range(0,3) for low and range(3,13) for medium. Same goes for high! – Merlin1896 May 09 '17 at 19:58

score 2 · Answer 2 · answered May 07 '17 at 21:53

2

I recommend you to use the plt.hist() function 3 times, each with a different color. You can set the range of each histogramm using the range parameter of the function. The legend is genereated by using the label parameter followed by a subsequent call of plt.legend().

answered May 07 '17 at 21:53

Padix Key

857
6
15

thanks. I'm a complete beginner here, is there anyway you could provide an example for this? – warrenfitzhenry May 09 '17 at 11:01

score 2 · Answer 3 · answered May 14 '20 at 10:14

If you want to color specific divisions with specific colors and label them accordingly you can use the following code:

import matplotlib.pyplot as plt
import numpy             as np
import seaborn as sns; sns.set(color_codes=True)

number_of_bins = 20
N, bins, patches = plt.hist(np.random.rand(1000), number_of_bins, rwidth=0.8)

#Define the colors for your pathces (you can write them in any format):
colors    = [(0, 0, 0), "b", "#ffff00", "red"]
#Define the ranges of your patches:
divisions = [range(1), range(1, 9), range(9, 14), range(14, 20)]
#If you want to label the regions/divisions:
labels    = ["Black", "Blue", "Yellow", "Red"]

#for each division color the parches according to the specified colors:
for d in divisions:
    patches[list(d)[0]].set_label(labels[divisions.index(d)])
    for i in d:
        patches[i].set_color(colors[divisions.index(d)])


plt.title("Plot Title")
plt.xlabel("X label")
plt.ylabel("Y label")
plt.legend(title="Legend Title")

score 1 · Answer 4 · answered Nov 20 '17 at 12:15

A more readable general solution without using the cmap as you want just 3 colors for specific intervals.

n, bins, patches = plt.hist(df1['Average'], 30)

for c, p in zip(bins, patches):
    if c > 0 and c < 4:
        plt.setp(p, 'facecolor', 'green')
    elif c >= 4 and c < 14  :
        plt.setp(p, 'facecolor', 'blue')
    else c>=14:
        plt.setp(p, 'facecolor', 'yellow')

plt.show()

Matplotlib/seaborn histogram using different colors for grouped bins

4 Answers4

Linked