hugo_clean_data: Clean your data

In pbiecek/hugo: I'm your personal assistant!

Description

This function fills missing values - a median for numeric variables and a mode for categorical variables (factors). Additionally, the outliers from numeric variables are replaced according to the IQR rule for outliers. In factors rare levels are merged into 'Other' level.

Usage

hugo_clean_data(data, prop = 0.01, num_to_fac_amount = 5)

Arguments

data	data.frame to clean
prop	proportion of occurence of the level in a categorical variable which decides which levels are rare
num_to_fac_amount	numeric columns with less than num_to_fac_amount unique values are treated as factors. You can disable this option by setting this parameter to 0

Value

data.frame that has been cleaned

Author(s)

Eliza Kaczorek

Examples

## Not run: 
# Dataset in base R: airquality
# There are 44 missing values
sum(is.na(airquality))

hugo_clean_data(airquality)
# The data was cleaned.

# Two original rows from data:

# Ozone Solar.R  Wind Temp Month Day
#     8      19  20.1   61     5   9
#     NA      NA 14.3   56     5   5

# After cleaning:

# Ozone Solar.R  Wind Temp Month Day
#     8      19 17.65   61     5   9
#  31.5     205 14.30   56     5   5

# We can see that the outlier in 'Wind' was
# replaced by the value Q3+1.5*IGR for this column.
# Missing values were replaced with medians.

## End(Not run)

pbiecek/hugo documentation built on May 12, 2019, 6:24 p.m.