SRM checker

Module is built for checking dataset for Sample Ratio Mismatch.

Inputs:

• df (required) - pandas data frame with experimental data, including group names (required) and dimenstions (optional)
  Example:

|    | group        | country   | platform   |   payment |
|---:|:-------------|:----------|:-----------|----------:|
|  0 | experimental | DE        | ios        |         0 |
|  1 | control      | IT        | web        |         1 |
|  2 | experimental | UK        | android    |         0 |
|  3 | experimental | IT        | web        |         1 |
|  4 | control      | DE        | android    |         0 |

• groups (required) - name of the column that contains groups names
• proportion (required) - dictionary of group names and related proportion in the dataset
  Example:

{
    "experimental": 0.5, 
    "control": 0.5
}

• dimensions (optional) - list of dimensions (column names in df) to check SRM by.

Exampe of usage:

check_srm(
        df=df, 
        groups='group', 
        proportion={'experimental':0.5, 'control':0.5}, 
        dimensions=['country', 'platform'])

Example of output:

|    | split    |   p value |
|---:|:---------|----------:|
|  0 | groups   |     0.035 |
|  1 | country  |     0.114 |
|  2 | platform |     0.714 |