Abstract:
Employee absenteeism creates a major operational and financial burden for organizations,
especially in labor-intensive sectors. This study aims to identify patterns and key factors
influencing employee absenteeism using a real-world dataset comprising 12,097 absence
records from 1,000 employees over a 2.5-year period. The dataset underwent rigorous pre-
processing to ensure data consistency and exploratory data analysis (EDA) applied to
uncover seasonal trends, demographic influences, and team-level variations using Python-
based analytical tools. Correlation analysis examined relationships between categorical and
numerical variables, with findings derived from the actual dataset. Results showed
absenteeism peaks with a cyclical components, with sick leave and annual leave marking the
highest number of absence records. Gender wise analysis indicated that females had higher
absence frequencies, mainly due to maternity leave, while employees aged 36-45 recorded
the majority of absence records. The study revealed that the average absence duration is 3.67
working days, and the mean tenure of employees is about 7.13 years. Correlation analysis
further revealed significant relationships among variables such as age, gender, tenure,
marital status, shift, and team, providing insights into workforce structure and patterns. The
results also show higher absenteeism among certain teams, illness-related absences, and
short-term absences, and the effect of demographic factors. Overall, these findings can guide
organizations in leveraging predictive analytics in workforce management, effective policy
implementation, and enhance operational efficiency, which is aimed at reducing
absenteeism.
