3

Suppose I have data [1,2,3, 7,8,9,9, 20,30,40,100,1000] for which I want to draw a histogram with Python. All the bins I care about is [0,5], [5,10], and [10, +∞). How can I do it?

The following wouldn't do it, of course.

import matplotlib.pyplot as plt

data = [1,2,3, 7,8,9,9, 20,30,40,100,1000]
plt.figure()
plt.hist(data, bins=5, color="rebeccapurple")
plt.show()
Paw in Data
  • 1,262
  • 2
  • 14
  • 32

1 Answers1

2

In case of forcing to show a histogram with customized x-range, you might need to process your data first.

I made a list of ranges and also x_ticklabels to show x-axis with range.

import matplotlib.pyplot as plt
import numpy as np

data = [1,2,3, 7,8,9,9, 20,30,40,100,1000,500,200]
data = np.array(data)
bin_range = [
    [0, 5],
    [5, 10],
    [10, 10000] # enough number to cover range
]

data2plot = np.zeros(len(bin_range))

for idx, (low, high) in enumerate(bin_range):
    data2plot[idx] = ((low <= data) & (data < high)).sum()
    
fig = plt.figure()
ax = fig.add_subplot(111)
ax.bar(range(len(bin_range)), data2plot)

x_labels = [
    f"{low}~{high}" for idx, (low, high) in enumerate(bin_range)
]

ax.set_xticks(range(len(bin_range)))
ax.set_xticklabels(x_labels)
plt.show()

enter image description here

ilkyu tony lee
  • 103
  • 8
  • Wow... this is bigger job than I'd expect! But using barplot is clever – Paw in Data Jul 15 '20 at 06:00
  • @PawinData Most sophisticated job is changing x-axis labels... without that, the code can be clear and short. – ilkyu tony lee Jul 15 '20 at 06:04
  • 1
    Just using `ax.set_xticklabels` without `ax.set_xticks` can be tricky, as they are set automatically. Better set `ax.set_xticks(range(len(bin_range)))` and `ax.set_xticklabels(x_labels)` without inserting the empty strings. – JohanC Jul 15 '20 at 10:07
  • @JohanC thx. Your method is so nice. Updated answer. – ilkyu tony lee Jul 16 '20 at 02:16