The students' approach
The students started by obtaining key data from Altria, including historical staffing information. They organized it by department, shift, job title and date to create a database of historical unplanned employee absences.
The team used four data sources provided by the company to create the database:
- Production calendar
- Cost center
- Employee absentee data
The team then used business analytics software to create a predictive model that could forecast unplanned absences by day of the week, shift and job title.
The forecasting model included three data exploration factors:
- Absentee rate by job class, shift, day of week and month
- Absenteeism by hire date
- Standard deviation in absentee count by month, year, and job class