Employee Retention Analysis
Salifort Motors seeks to improve employee retention and gain insights into the factors that influence employee attrition the most.
Our solution involves building a binary classification model that predicts which employees will leave. The XGBoost classifier performed best with an accuracy of 98.2%
KEY INSIGHTS:
-
Employees at this organization seem to be overworked, with an employee working ~40 hours more than is expected from a 9 to 5 job. They should consider placing a cap on the number of hours a staff can work weekly.
-
Standardize the criteria for promoting employees.
-
Cap the number of projects an employee can participate in.​
-
The majority of employees leave in their third to fifth year, with those who leave in their fourth year having very low satisfaction scores.
​
Click on the link for notebook file.
​​
Retention by Satisfaction level & Tenure
-
Employees who left in their 4th year had very low satisfaction levels. This could be because they expected a promotion and didn't get one.
-
Those that have stayed more than 6 years in the company don't leave.
Retention by Satisfaction level
-
Retained Staff have a higher satisfaction level on average when compared to those who left.
Retention by Monthly hours worked and No of Projects
-
Employees with only 2 projects have a high attrition rate and work very few hours. This is probably because it is already anticipated that they will be leaving any time soon.
-
All employees on 7 projects leave and those who are assigned to more than 4 projects work significantly higher hours.
