StandardDeviation Category for groovy, inspired by Advanced Rails

StandardDeviationCategory.groovy

import static java.lang.Math.*
import static java.math.RoundingMode.CEILING

class StandardDeviationCategory {

    static Double mean(Collection list) {
        list.with{
            scale(sum() / size())
        }
    }

    static Double stdev(Collection list){
        scale(sqrt(list.collect{ num -> pow(num - list.mean(),2) }.sum() / (list.size() -1)))
    }

    static String confidenceInterval(Collection list) {
        list.with{
            confidenceInterval(list, mean(), 2*stdev())
        }
    }

    static String confidenceInterval(Double mean, Double squireStdev) {
        "${scale(mean - squireStdev)}..${scale(mean + squireStdev)}"
    }

    static String stat(Collection list){
        list.with{
            def mean = mean()
            def stdev = stdev()
            "mean: ${mean} " +
            "standard deviation: ${stdev} " +
            "confidence interval: ${confidenceInterval(mean, 2*stdev)}"
        }
    }

    static Double scale(BigDecimal value){
        value.setScale(3, CEILING)
    }
}

just a little suggestion - line 13 has list.mean() inside the closure, this will get called each iteration of the loop. Calculate it outside the closure and use that value. much faster

I'm wondering why you subtract 1 from list.size() in the std dev formula but not for mean.

I'm wondering why you subtract 1 from list.size() in the std dev formula but not for mean.

Just in case anyone else stumbles across this and is wondering, the reason is that is the definition of the standard deviation: https://en.wikipedia.org/wiki/Standard_deviation#Basic_examples