healthcareai (version 1.2.4)

findVariation: Find high variation

Description

Search across subgroups and surface those that have coefficient of variation * volume above a particular threshold

Usage

findVariation(df, categoricalCols, measureColumn, dateCol = NULL,
  threshold = NULL, wideOutput = TRUE)

Arguments

df

A dataframe

categoricalCols

Vector of strings representing categorical column(s)

measureColumn

Vector of strings representing measure column(s)

dateCol

Optional. A date(time) column to group by (done by month)

threshold

A scalar number, representing the minimum impact values that are returned

wideOutput

If TRUE (default) categoricalCols and measureColumn will be mixed in with results. If FALSE, output table will have Group and Measure columns that reflect categoricalCols and measureColumn, and results will appear in their own columns.

Value

A dataframe of eight columns. MeasureVolumeRaw denotes number of rows in the particular subgroup; MeasureVolumePercent denotes percent of rows in that subgroup as a percentage of the above subgroup (i.e., F within Gender); MeasureImpact is the subgroup COV * VolRaw (i.e., num of rows).

References

http://healthcareai-r.readthedocs.io

See Also

healthcareai calculateCOV createVarianceTallTable

Examples

# NOT RUN {
df <- data.frame(Dept = c('A','A','A','B','B','B','B','B'),
                 Gender = c('F','M','M','M','M','F','F','F'),
                 LOS = c(3.2,NA,5,1.3,2.4,4,9,10))

categoricalCols <- c("Dept","Gender")

dfRes <- findVariation(df = df, 
                       categoricalCols = categoricalCols,
                       measureColumn = "LOS")

dfRes
# }