Add Outlier Flags Based on 5-MAD Bounds

add_outlier5std_column calculates and appends outlier flags for specified indicators in a dataset. An outlier is defined as a value falling outside the range defined by the median ± 5 times the Median Absolute Deviation (MAD). This method is applied on a per-group basis, allowing flexibility for grouping by columns like district.

Usage

add_outlier5std_column(.data, indicators, group_by = "district")

Arguments

.data

A cd_data tibble containing health indicator data.

indicators

A character vector specifying the names of the indicator columns to analyze for outliers.

group_by

A character string or vector specifying the column(s) to group by when calculating the median and MAD (default is 'district').

Value

A tibble with additional columns for each indicator, named {indicator}_outlier5std, containing a value of 1 if the observation is an outlier and 0 otherwise.

Details

• Median and MAD Calculation: For each indicator, the median and MAD are calculated using data from years prior to the most recent year in the dataset. The last year is excluded to prevent potential contamination of the baseline statistics.

• Outlier Definition: A value is flagged as an outlier if it falls outside the range: Lower Bound = median - 5 * MAD Lower Bound = median + 5 * MAD

• Grouping: Calculations are performed separately for each group specified by the group_by parameter.

• Generated Columns: For each indicator, the function generates an outlier flag column named {indicator}_outlier5std. The column contains 1 if the value is an outlier and 0 otherwise.

See also

add_mad_med_columns() for computing and appending the median and MAD columns.

Examples

if (FALSE) { # \dontrun{
  # Add missing value flags for all indicators
  add_outlier5std_column(data, indicators = 'indicator1',
                                       group_by = 'district')
} # }