opal-analytics-statstest

Overview

This Opal contains a variety of statistical tests taken from both python and R stats packages. Below is a list of currently available packages:

• Wilcoxon signed-rank (paired)
• Wilcoxon rank sums (unpaired)

Installation

pip install git+https://github.com/Bedrock-py/opal-analytics-statstests.git

Parameters Spec for Wilcoxon

self.parameters_spec = [
    { "name" : "x_column", "attrname" : "x_column", "value" : "x.column_name", "type" : "input" },
    { "name" : "y_column", "attrname" : "y_column", "value" : "y.column_name", "type" : "input" },
    { "name" : "paired", "attrname" : "paired" , "value" : "False", "type" : "input"}
]

• x_column The reference to the column to be used for the x data Format is x.column_name or y.column_name
• y_column The reference to the column to be used for the y data Format is x.column_name or y.column_name
• paired True to use the signed-rank test, False to use the rank-sums test

Note that x_column or y_column can come from either input matrix. Prefix the column name with x. or y. to specify which column it refers to

Requires either two matrices or one matrix with both samples (The columns are selected from the parameters)

• x_data.csv The matrix with x data (Uses the matrix.csv from the input)
• x_features.txt A list of column names for the matrix (one name per row)
• y_data.csv The matrix with y data (Pass empty string "" or empty object "{}" if not required) (Uses the matrix.csv from the input)
• y_features.txt A list of column names for the matrix (one name per row) (Pass empty string "" if not required)

Outputs the following files

matrix.csv The Output of the wilcoxon test