Employee Retention and Flight Risk Modelling
Predicting student officer turnover to enhance workforce stability for a public sector organisation.
The Opportunity
A public sector organisation sought a data science partner to conduct a comprehensive analysis of data related to the retention of new Student Officers. The goal was to develop a predictive model to identify key variables contributing to turnover and predict which officers have a high probability of leaving, allowing for proactive intervention.
What we did
✔︎ Conducted a data audit to assess data quality and completeness ✔︎ Built a predictive model for Student Officer flight risk ✔︎ Developed an Excel tool for easy use and updates by by the organisation ✔︎ Provided recommendations for future data collection and usage ✔︎ Ensured adherence to data ethics guidelines
4Wks
Engagement to develop a predictive model.
How we did it
This project was a statistical modelling project aimed at understanding and predicting the retention of employees. Through collaboration with stakeholders from the organisation, we conducted a data audit to assess the quality and volume of existing data. We worked with Legal Services and IT Security teams to ensure GDPR compliance and best-practice data ethics. We recommended additional variables to enrich the dataset and defined an appropriate evaluation metric for the predictive model. We performed exploratory data analysis (EDA) to understand the data and identify patterns. Time series analysis was used to analyse trends, and hypothesis testing was conducted to evaluate claims about the data distribution. Geographic analysis was performed to uncover location-based insights. We built a logistic regression model to predict employee flight risk, ensuring transparency and explainability. The model coefficients were stored in an Excel tool, allowing the organisation to make predictions and understand the impact of different variables. We provided guidance on hiring the required skillsets for ongoing model maintenance.
