2.3 Understanding the Variables

The tables below describe each variable in the dataset, organized by category. Understanding these variables now will make it much easier to work with the data when we begin analysis in Chapter 4.

2.3.1 The Outcome Variable

Variable	Description	Type
Absenteeism time in hours	Total hours absent for this event	Numerical

This is our target variable — the thing we ultimately want to understand and predict.

2.3.2 Employee Demographics

Variable	Description	Type
ID	Unique identifier for each employee	Identifier
Age	Employee’s age in years	Numerical
Education	Highest level attained (1 = high school, 2 = graduate, 3 = postgraduate, 4 = master/doctorate)	Ordinal
Son	Number of children	Numerical (discrete)
Pet	Number of pets	Numerical (discrete)

2.3.3 Health Indicators

Variable	Description	Type
Body mass index	BMI calculated from height and weight	Numerical
Weight	Employee’s weight (kg)	Numerical
Height	Employee’s height (cm)	Numerical
Social drinker	Whether the employee drinks socially (1 = yes, 0 = no)	Binary
Social smoker	Whether the employee smokes socially (1 = yes, 0 = no)	Binary

2.3.4 Work Factors

Variable	Description	Type
Transportation expense	Monthly transportation cost (Brazilian Real)	Numerical
Distance from residence to work	Commute distance in kilometers	Numerical
Service time	Years of employment at the company	Numerical
Work load average/day	Average daily workload (target units)	Numerical
Hit target	Percentage of productivity target achieved	Numerical
Disciplinary failure	Whether the employee has a disciplinary record (1 = yes, 0 = no)	Binary

2.3.5 Absence Context

Variable	Description	Type
Reason for absence	Coded reason using ICD categories (see below)	Categorical
Month of absence	Month when the absence occurred (1–12)	Categorical
Day of the week	Day of absence (2 = Monday through 6 = Friday)	Categorical
Seasons	Season of absence (1 = summer, 2 = autumn, 3 = winter, 4 = spring)	Categorical

A Note on Data Types

You may have noticed that the variables above are labeled as different types — numerical, categorical, ordinal, and binary. These distinctions matter because the type of data determines which analytical techniques are appropriate. Numerical data (like age or BMI) can be averaged and used in mathematical calculations. Categorical data (like season or day of the week) represents groups or categories. Ordinal data (like education level) is categorical but has a natural order. Binary data (like social drinker) has only two possible values. We will revisit these distinctions throughout the book as we select appropriate methods for each type of variable.