histogram

Generates histograms similar to R’s hist and numpy’s histogram functions. The interface is relatively stable and a decent test/spec suite is in place, but please consider this alpha software. Inspired by Richard Cottons’s matlab implementation and the wikipedia histogram article.

Examples

Typical usage:

require 'histogram/array'  # enables Array#histogram

data = [0,1,2,2,2,2,2,3,3,3,3,3,3,3,3,3,5,5,9,9,10]
# by default, uses Freedman-Diaconis method to calculate optimal number of bins
# and the bin values are midpoints between the bin edges
(bins, freqs) = data.histogram 
# equivalent to:  data.histogram(:fd, :tp => :avg)

Multiple types of binning behavior:

# :fd, :sturges, :scott, or :middle  (median value between the three methods)
data.histogram(:middle)
(bins, freqs) = data.histogram(20)                         # use 20 bins
(bins, freqs) = data.histogram([-3,-1,4,5,6])              # custom bins

# bins are midpoints, but can be set as minima
(bins, freqs) = data.histogram([-3,-1,4,5,6], :tp => :min) # custom bins with :min

# can also set the bin_width (which interpolates between the min and max of the set)
(bins, freqs) = data.histogram(:bin_width => 0.5)

Multiple Datasets:

Sometimes, we want to create histograms where the bins are calculated based on all the data sets. That way, the resulting frequencies will all line up:

# returns [bins, freq1, freq2 ...]
(bins, *freqs) = data.histogram(30, :other_sets => [[3,3,4,4,5], [-1,0,0,3,3,6]])

Histograms with weights/fractions:

# histogramming with weights (uses the second array for weights)
w_heights = [data, [3,3,8,8,9,9,3,3,3,3]]
w_heights.histogram(20)

Works great with NArray, too:

require 'histogram/narray'  # enables NArray#histogram
# if the calling object is an NArray, the output is two NArrays:
NArray.float(20).random!(3).histogram(20)
# => [bins, freqs]  # are both NArray.float objects

Installation

gem install histogram

See Also

aggregate, rsruby