Analysis of Employee Attrition: Insights from an HR Analytics Endeavour

Employee attrition rate measures employee turnover—the number of people who leave a company over a given period for various reasons. It is calculated as a percentage of separated employees to the average number of employees in that company.

Aim of the Project

The essence of this project was to aid a company in understanding the attrition in their firm. An appreciation of the factors surrounding attrition would help the firm to minimize the attrition rate, so as to retain staff for longer, and minimize turnover cost. It was to explore the diverse features contained in the dataset, how they relate to each other, and how they bear relevance to the company’s attrition. Since I had earlier done a project in Excel, it was fairly easier for me to work my way around the spreadsheet especially to create the visualisations and dashboard.

Project Overview:

The project was an exploratory data analysis of a company’s attrition dataset, segmenting the company’s attrition count by the employees’ demographic features in the dataset.

Data Collection, Inspection, and Cleaning:

The dataset I used was in CSV format which I imported into my spreadsheet. The dataset contained various features centred around the employees’ details, demographics and attrition count. Some of the features include Gender, Marital Status, Age, Education, Job Role, Department, Job Satisfaction, Total Working Years, etc. I inspected the dataset and it was clean; there were no missing values or outliers. After this, I went through the data extensively to understand what each variable meant, and its relevance to the project.

Feature Engineering and Enhancement:

Like my previous Excel project, I created a column, Age_Band, that groups the employees into corresponding age groups of Under 25, 25-34, 35-44, 45-54 and Over 55 using the Nested IF function.

Pivot Tables and Summaries:

On a new sheet, I created pivot tables using the relevant columns from the dataset; I also created visualisations to correspond to each pivot table. The pivot tables were to bring out summaries of the data pertinent to employee attrition. The pivot tables contained summaries of Count of employees, sum of attrition count, average age, average job satisfaction, attrition by education, job roles and department. I also calculated the attrition rate and active employee count.

Visualizations and Insights:

For visualisation, I employed a wide range of charts and graphs to represent the pivot tables and the employees’ KPIs. The visualisation helped the analysis a lot by making it easy to understand what the data and pivot tables say at a glance. I then created a new sheet where I brought all these visualisations together to form a beautiful dashboard. I also added filters and slicers to enhance the visualisation and give more detail about specific variables the data.

Analysis Findings:

A lot of interesting insights were revealed from this exploratory data analysis; the first of which was that the company had an attrition rate of 16.12%—this is higher than the generally accepted healthy attrition rate of 10%. Furthermore, on a scale of 4, the average job satisfaction rating is 2.6. Most of the separated employees are in the age band 25-34, who have a Bachelors degree, and are male Laboratory technicians. These are just a few of the insights gathered from the analysis.

Conclusion:

By reason of this analysis—and exploratory data analyses, in general—companies are able to have a firmer grasp on the less obvious, but just-as-important factors that influence employee attrition. With this information, accurate, data-driven conclusions and decisions can be made to bring the company’s attrition to a healthy rate.