Python script to plot a histogram of a set of numbers.

histogram.py

#!/usr/bin/env python

'''
This script reads a set of numbers and plots a histogram.
'''

import sys
import matplotlib.pyplot as plt
import numpy as np

if len(sys.argv) != 2:
    print 'Usage:  python {} path/to/data/file'.format(sys.argv[0])
    print
    print 'Or:  pbpaste | python {} stdin'.format(sys.argv[0])
    print '  (This way it will plot whatever you most recently copied.)'
    print
    print '(The data should be a sequence of numbers separated by either "," or newlines.)'
    print
    print 'Error: This program takes one argument. See above.'
    print 'Quitting.'
    sys.exit()

filepath = sys.argv[1]

if filepath == 'stdin':
    data_from_file = sys.stdin.read().strip()
else:
    data_from_file = open(filepath).read().strip()

def is_delim_by(delim):
    return len(data_from_file.split(delim)) > 3

if is_delim_by(',') and '\n' not in data_from_file:
    numbers = np.array([float(x) for x in data_from_file.split(',')])
elif is_delim_by('\n') and ',' not in data_from_file:
    numbers = np.array([float(x) for x in data_from_file.split('\n')])
else:
    print 'Error: Not sure how to parse the data.'
    print 'The data should be a sequence of numbers separated by either "," or newlines.'
    print 'Quitting.'
    sys.exit()

p95 = np.percentile(numbers, 95)
p05 = np.percentile(numbers,  5)

plt.hist(numbers, bins=50, range=(p05, p95))
plt.title('Histogram')
plt.xlabel('Value')
plt.ylabel('Frequency')
plt.show()