Histograms are great for exploring data, but numpy and matplotlib are heavy and overkill for quick analysis. They also can't be easily used on remote servers over ssh. Don't even get me started on installing them.
Bit.ly's data_hacks histogram.py is
great but difficult to use from python code directly (it requires an
optparse.OptionParser to pass histogram options). This is histogram.py
repackaged for convenient script use.
Python 3.11.3 (tags/v3.11.3:f3909b8, Apr 4 2023, 23:49:59) [MSC v.1934 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> from text_histogram import histogram
>>> histogram([1, 2, 2, 2, 2, 3, 3, 3], graph_char='#')
# NumSamples = 8; Min = 1.00; Max = 3.00
# Mean = 2.250000; Variance = 0.437500; SD = 0.661438; Median 2.000000
# each # represents a count of 1
1.0000 - 1.2000 [ 1]: #
1.2000 - 1.4000 [ 0]:
1.4000 - 1.6000 [ 0]:
1.6000 - 1.8000 [ 0]:
1.8000 - 2.0000 [ 4]: ####
2.0000 - 2.2000 [ 0]:
2.2000 - 2.4000 [ 0]:
2.4000 - 2.6000 [ 0]:
2.6000 - 2.8000 [ 0]:
2.8000 - 3.0000 [ 3]: ###
>>> histogram([1, 2, 2, 2, 2, 3, 3, 3], graph_char='#', display_empty_buckets=False)
# NumSamples = 8; Min = 1.00; Max = 3.00
# Mean = 2.250000; Variance = 0.437500; SD = 0.661438; Median 2.000000
# each # represents a count of 1
1.0000 - 1.2000 [ 1]: #
1.8000 - 2.0000 [ 4]: ####
2.8000 - 3.0000 [ 3]: ###
pip uninstall text_histogram text_histogram3 # avoid conflicts
pip install text_histogram32Source: https://github.com/clach04/text_histogram32
Fork of https://github.com/Kobold/text_histogram with a few tweaks:
- Same name space as original text_histogram
- Python 3 and 2.7 support
- Support for control over character used for the bar (see graph_char and DEFAULT_graph_char option)
- Support for NOT displaying empty buckets/bins/intervals (see display_empty_buckets and DEFAULT_display_empty_buckets option)
- Fixes
- Kobold#4 - Zero for min or max value
- when min == max value
- improved error handling for empty data sets
ToDo items
- Test suite
- see https://github.com/piccolomo/plotext/blob/master/readme/bar.md#histogram-plot
- see https://github.com/nschloe/termplotlib?tab=readme-ov-file#similar-projects
- Improve dynamic sizing of string and bar
- Support for non-linear buckets/bins/intervals (see display_empty_buckets option)?
- Support for Scott's rule https://en.wikipedia.org/wiki/Scott%27s_rule ?
- Bug? Kobold#2 bitly/data_hacks#22
- Bug different results for python 2 (missing bar) compared with Python 3 - test case below:
from text_histogram import histogram python2_3_difference_with_low_counts = [] python2_3_difference_with_low_counts += [ 2540] * 1142 python2_3_difference_with_low_counts += [ 5071] * 163 python2_3_difference_with_low_counts += [ 7602] * 67 python2_3_difference_with_low_counts += [ 10134] * 28 python2_3_difference_with_low_counts += [ 12665] * 17 python2_3_difference_with_low_counts += [ 15196] * 14 python2_3_difference_with_low_counts += [ 17728] * 4 python2_3_difference_with_low_counts += [ 20259] * 4 python2_3_difference_with_low_counts += [ 22790] * 2 python2_3_difference_with_low_counts += [ 25322] * 4 python2_3_difference_with_low_counts += [ 27853] * 2 python2_3_difference_with_low_counts += [ 30384] * 4 python2_3_difference_with_low_counts += [ 32915] * 2 python2_3_difference_with_low_counts += [ 40509] * 2 python2_3_difference_with_low_counts += [ 45572] * 1 python2_3_difference_with_low_counts += [ 50635] * 1 python2_3_difference_with_low_counts += [ 55697] * 1 python2_3_difference_with_low_counts += [ 63291] * 1 python2_3_difference_with_low_counts += [129105] * 1 python2_3_difference_with_low_counts += [162012] * 1 python2_3_difference_with_low_counts += [253139] * 1 print(python2_3_difference_with_low_counts) histogram(python2_3_difference_with_low_counts, buckets=100, graph_char='#', display_empty_buckets=False)