带有柱状图的直方图值的百分比？

Question

I am creating a histogram in python and I want the bin edges to be a percentage of given values (5-10%). What would be the best way to go about this so that I don't leave gaps in the bin boundaries, and don't have to pre-set some values for the bin boundary calculation?

Answer 1

In general, it's convenient to create histograms using pre-defined tools like numpy.histogram , though your newly posted comment- suggesting that you're using matplotlib- is also totally fine. Either way allows you to create a set number of automatically determined bins of equal width...

import numpy
data = [0,1,1,1,1,1,1,2,3,3]
hist, edges = numpy.histogram( data , bins = 10)
>>> hist
array([1, 0, 0, 6, 0, 0, 1, 0, 0, 2])
>>> edges
array([ 0. ,  0.3,  0.6,  0.9,  1.2,  1.5,  1.8,  2.1,  2.4,  2.7,  3. ])

...Or, in the odd case where you want predefined bins (possibly of different width), you can specify the bin edges yourself (read the docs for information on how this works):

>>> hist, edges = numpy.histogram( data , bins = [0,.5,1., 1.5,2,3])
>>> hist 
array([1, 0, 6, 0, 3])
>>> edges
array([ 0. ,  0.5,  1. ,  1.5,  2. ,  3. ])
>>> 

Just be careful about using drastically different bin sizes , however. In many cases this sort of coarse graining could distort the relationships between the numbers you're trying to compare.

As for your value +/-10% boundary?

preferred_bin_centers = [0,1,2,3]
bin_pairs = [ ( 0.9* v , 1.1*v ) for v in preferred_bin_centers ]
>>> [[0.0, 0.0], [0.9, 1.1], [1.8, 2.2], [2.7, 3.3000000000000003]]

Or, flattened into a list form that could be used by numpy.histogram...

bin_edges = sum( [  [ 0.9* v , 1.1*v ]  for v in values ]    , [] )

>>> [0.0, 0.0, 0.9, 1.1, 1.8, 2.2, 2.7, 3.3000000000000003]

(Note from the first two items of the above list that this code gives confusing bin edges if one of your bin centers is 0; I left that in solely as an example of what to watch out for)

Incidentally, the bin edges as defined above will also create intermediate bins outside your desired range. For example, if you bin items within +/- 10% of 1,2, and 3, then inherently, there will also be a bin between 2.2 and 2.7 (the "outside edges" of your desired bins) where numbers like 2.5 would go. If you have values that exist in between your desired bins, then you may want to adjust your cutoffs or visualization accordingly.

Answer 2

Maybe I'm oversimplifying your question?

def bins(data, nbins):
    range = max(data) - min(data)
    binsize = range / float(nbins)
    bins = [x * binsize for x in range(nbins)]