cancel
Showing results for
Did you mean:

# alteryx Use Cases

Learn how you can leverage Alteryx in your organization.
Announcement | Are you an advocate for Alteryx? Check out our Advocacy.Amplified program to get recognized for your evangelism!

# Predicting Employee Attrition Rate using Alteryx Predictive Suite

Bolide
Created on

Overview of Use Case

Michael Utama built a workflow on Alteryx Designer to predict the rate of employee attrition. Using the predictive tools and plugging in a sample of historic Human Resources (HR) data, he discovered that the length of an employee’s commute was a more important factor than salary or promotion in considering whether to leave a job or not. Michael hopes that any person in HR could use his solution to help companies retain talent and save money.

Describe the business challenge or problem you needed to solve
• HR teams are often data rich but insight poor. Despite having access to a data gold mine, too many HR teams spend most of their time on admin tasks or legal issues.
• Employee retention: Retaining key employees is critical to the long-term health and success of any business. It costs money to replace employees and takes time to train them, delaying projects and deadlines.

Basic Exploratory Data Analysis (EDA)

We observed that variables such as monthly income are highly skewed to the left (mode of approx. \$5000 monthly salary, with very few employees earning above \$10,000. This is sensible because there can only be that many Director, C-suite level employees).

Check Correlation between Variables

Tease out the variables that have high correlation (r > 0.5) with one another.

Predictive Modelling

• Remove variables with high correlation, e.g. [YearsAtCompany]
• Correct Class Imbalance with Undersampling of [Attrition] = '0'
• Noticed that [MonthlyIncome] is highly skewed, we Log-Transform this variable to correct this over-skewness and normalize the variable's range of values
• The initial modelling. Conduct stepwise regression to select the more statistically significant variables to increase AUC of predictive model

• Ascertaining which variables exert considerable statistically significant effects on this company's Attrition Rate (we sort by the ascending order of the logit).
• If coefficient (logit) is positive, the effect of this predictor (on attrition rate) is positive & vice versa.

• Ascertaining which variables exert considerable statistically significant effects on this company's Attrition Rate (we sort by the ascending order of the logit).
• If coefficient (logit) is positive, the effect of this predictor (on attrition rate) is positive & vice versa.

Describe the benefits you have achieved
Any industry and any person can use this solution. Alteryx’s ease of use allow anyone to run this workflow and understand the output. “It’s like kids playing with Legos,” Michael said.

Related Resources

See samples data, workflow and macros below.