Home

CompStatsPython is a project to release some old statistics code into the wild. I've been developing this for fun for a few years and it's wasted here. It's also unit tested (check if a test exists to see if a routine's been tested) so the algorithms should be reliable. Finally, it forms the basis of code for a book I'm releasing about performing statistics with Python.

The list of planned tests is below, but many more (especially inferential tests) will be included.

• VSort - sort a vector
• MSort - sort a numpy.matrix
• CalculateRanks - for calculating the ranks of a numpy.matrix
• GetSSCP_M - calculates the sum of squares and cross-products numpy.matrix
• GetVarsCovars_M - calculates the variances and covariances numpy.matrix
• GetVariances - calculates the variances of a numpy.matrix of variables
• GetStdDevs - calculates the standard deviations of a numpy.matrix of variables
• GetCorrelationMatrix - calculates the correlation numpy.matrix
• Count - returns the number of non-missing data
• Sum - returns the sum of non-missing data
• Minimum - returns the minimum of non-missing data
• Maximum - returns the maximum of non-missing data
• Range - maximum minus the minimum
• Proportions - Proportions of values
• relfreqmode - Relative frequency of the mode
• CumSum - Cumulative sum
• CumProduct - Cumulative product
• CumPercent - Cumulative percentage
• Frequencies - Frequencies of values
• TrimmedData - Trims data
• TrimmedMean - Trimmed mean
• BiTrimmedMean - Trimmed mean with different trims each end
• Mean - Arithmetic mean
• Median - Median
• Mode - Modal value(s)
• Moment - Moments of the distribution
• TukeyQuartiles - returns Tukey's hinges
• MooreQuartiles - returns Moore & McCabe's hinges
• SPQuantile - quantile used by S-Plus
• TradQuantile - quantile used by SPSS
• MidstepQuantile - mid-step qua
• Q1 - Q1 quantile from Hyndnumpy.man & Fan
• Q2 - Q2 quantile from Hyndnumpy.man & Fan
• Q3 - Q3 quantile from Hyndnumpy.man & Fan
• Q4 - Q4 quantile from Hyndnumpy.man & Fan
• Q5 - Q5 quantile from Hyndnumpy.man & Fan
• Q6 - Q6 quantile from Hyndnumpy.man & Fan
• Q7 - Q7 quantile from Hyndnumpy.man & Fan
• Q8 - Q8 quantile from Hyndnumpy.man & Fan
• Q9 - Q9 quantile from Hyndnumpy.man & Fan
• InterquartileRange -
• SS - sum of squares
• SSDevs - sum of squared deviations from the mean
• SampVar - sample variance
• PopVar - population variance
• SampStdDev - sample standard deviation
• PopStdDev - population standard deviation
• StdErr - standard error
• CoeffVar - coefficient of variation
• ConfidenceIntervals - returns the confidence intervals
• numpy.maD - Median absolute deviation
• GeometricMean - the geometric mean
• HarmonicMean - the harmonic mean
• MSSD - mean of subsequent squared deviations
• Skewness - returns the skewness
• Kurtosis - returns the kurtosis
• StandardScore - transforms a vector into a standard (ie, z-) score
• EffectSizeControl - returns an effect size if a control condition is present
• EffectSize - returns an effect size if no control is present
• FiveNumber - Tukey's five number sumnumpy.mary (minimum, lower quartile, median, upper quartile, maximum)
• OutliersSQR - returns two arrays, one of outliers defined by 1.5 * IQR, and the other without these outliers

Finally, there's also a small but under-development library for statistics with Prolog! This is a side project I'm doing for more numerical fun.

Have fun