Turnover Reduction With Machine Learning

Overview 🔎

Retaining talent can be a challenging task for any company. Analyzing employer data, I could identify key conflict points and specific situations that were getting in the way of maintaining a good team together for the long run.

This project was developed to help my chances of landing a job at Ernst & Young. They were looking for someone experienced in Machine Learning for HR. I didn't have any professional experience, but I figured a related project in my portfolio would help.

Unfortunately, by the time I finished the project, the position had been filled. Not a problem, since it was a great opportunity for me to expand my range and work on something different for my portfolio.

Problem & Solution 🤝

The dataset used for this project was created and distributed by IBM Data Scientists, to provide other data scientists with the data to come up with innovative solutions in this area.

My main goal was to identify problems and conflicts, that could influence the retention of employees. Some of the main goals of the project: • Find and understand points of attrition between employees. • Understand the employees of the company, regarding their experience, educational level, tenure, and more. • Understand how much of a role gender plays in the company.

Process 🛣

Through analysis, I was able to find out some very interesting insights.

To start, gender-wise, women received better salaries when compared to men, in every department, and at almost every level of education and hierarchy within the company. The only part of the company where men got paid more was at the very top of the salary range.

It was also interesting to find that people who had been recently promoted showed a higher level of attrition, which may indicate that the company needs to improve its transition process when employees get relocated or promoted.

Results 🎁

At the end of the project, I was able to identify important attrition points that could be avoided with simple actions. Taking these actions could result in a reduction in employee turnover in the company.

Takeaways 📣

Of course, the data was fabricated and there is a chance that the gender equality factor was intended by the creators of the dataset. However, I found it very interesting to work on this aspect of the project.

It was also surprising to see how much can be achieved by simply taking a step back and analyzing the data. Using Data Science in HR might not come across as obvious, but it's an incredible opportunity that businesses should take seriously.