Online Program

Modeling and Optimizing Job Retention: A Case Study (304201)

Keywords: Data Science, Machine Learning, Case Study, Model Implementation, Data Management

Job placement and long term retention for chronically unemployed people has always been a challenge for state and national governments. A global service provider tasked with placing and retaining job seekers for a government entity partnered with the Statistical Applications and Innovations Group (SAIG) at Virginia Tech on a project to improve job retention metrics. Our goal was to use statistics and machine learning methods to predict which jobs and services would maximize retention. In this case study we will cover technical, organizational and business challenges in detail including using administrative data for prediction, data cleaning, and model fitting versus model implementation. Specifically, we will discuss the do’s and don’ts of housing large scale data across multiple databases and how to turn messy, administrative data into clean data ready for modeling. Technical details will be discussed within the confines of our confidentiality agreement with the client. Finally, we will focus on the importance of effective team communication and team decision making, particularly in the context of a project with a partner in the early phases of their analytics journey.