whossname / stream_stats

Concurrent calculation of count, mean and standard deviation

StreamStats

Enables concurrent calculation of count, mean and standard deviation. New values can be aggregated into an existing stat tuple and two stat tuples can be merged into one.

Inspired by the following article by John D. Cook: https://www.johndcook.com/blog/skewness_kurtosis/

Installation

The package can be installed by adding stream_stats to your list of dependencies in mix.exs:

def deps do
  [
    {:stream_stats, "~> 0.1.0"}
  ]
end

Example usage

Given two lists of numbers values_1 and values_2 the two lists can be aggregated independently, then combined into a single stats tuple:

  stream_1 = StreamStats.reduce(values_1)
  stream_2 = StreamStats.reduce(values_2)
  stats = StreamStats.combine(stream_1, stream_2)

  {count, mean, _m2} = stats
  std_dev = StreamStats.standard_deviation(stats)