AzureFeeds

Our Team (Sorted Alphabetically):

 Master Business Analysis & Modelling, Radboud University

Master Strategic Human Resources Leadership, Radboud University

Student Exchange at Radboud University

Project Overview

At Radboud University, our team in the course Data-Driven Analytics for Responsible Business Solutions, embraced a unique opportunity to apply our passion for data analytics to a real-world challenge. Tasked with analyzing employee turnover at VenturaGear—a company committed to fostering a thriving workplace—we conducted an in-depth study to uncover the root causes of attrition. By leveraging rigorous data analytics, machine learning techniques, and Power BI visualizations, we identified key drivers of employee satisfaction and retention. This blog outlines our approach, findings, and actionable strategies to address similar challenges.

Understanding the Turnover Challenge

VenturaGear, a company with diverse operational units, faced rising employee turnover, prompting our data analytics team to identify its root causes. Initial concerns highlighted external competition and internal workplace factors, but the broad question of “why are employees leaving?” required refinement for actionable insights. To tackle this, we first reviewed the dataset and its accompanying data manual to gain a comprehensive understanding of the available data. We then conducted exploratory data analysis on current employee records, focusing on key variables: organizational hierarchy (e.g., employee level), tenure, business units (e.g., inventory, manufacturing), pay frequency, pay rate, and performance metrics.

This approach led us to refine our research question: How do organizational hierarchy, tenure, business unit, pay structure, and performance metrics influence employee satisfaction and turnover at VenturaGear? By narrowing the scope, we could systematically explore these factors, uncovering patterns and drivers of attrition to set the stage for targeted recommendations.

Data Preparation

To answer our refined research question effectively, we needed to examine VenturaGear’s turnover challenge from multiple perspectives. Using Power BI, we created a comprehensive data model to avoid tunnel vision and ensure a holistic analysis. This began with a thorough review of the dataset to assess data quality, checking for inconsistencies, data type accuracy, and missing values (Hair et al., 2018). For instance, we converted the “EndDate” field in the HR_EmployeeDepartmentHistory table from text to date format and corrected errors, such as one employee’s “EndDate” preceding their “StartDate,” which was removed to maintain data integrity. We streamlined the analysis by importing only relevant tables from the Human Resources dataset, establishing relationships using the data dictionary and verifying cardinality to prevent errors (Arnold, 2022). 

To enhance our analysis, we developed several new columns and measures, including WorkingLifeTime, calculated with DATEDIFF to assess employee tenure, and ActiveCountTrue and ActiveRatioTrue to evaluate active employee statuses. We also categorized organizational levels with OrgLevelLabel to reflect VenturaGear’s hierarchy. To ensure accurate pay analysis, we implemented the Interquartile Range (IQR) method for rate outlier detection. While we flagged missing data in fields like BirthDate and HireDate, we left these unchanged to avoid making assumptions. This data preparation allowed us to conduct a thorough analysis of turnover drivers across various levels, departments, and pay structures.

Our Journey in Analysis

With a robust data model in place, our journey in analysis focused on translating VenturaGear’s employee data into actionable insights. We began by analyzing turnover within different units to understand the issue, then segmented employees by tenure, units, and levels using the MECE framework for a deeper understanding (Rasiel, 1999). Next, we measured well-being through job satisfaction, pay frequency, and rate, identifying markers hinting toward extrinsic motivation. To quantify these influences, we developed a machine learning model using Microsoft Fabric AutoML, treating satisfaction as a continuous variable (Müller & Guido, 2016). The model’s feature importance analysis revealed that pay frequency (0.55) and organizational level (0.26) were the strongest predictors, while pay rate and shift ID had minimal impact.

However, the trained regression machine-learning model’s prediction performance (R² of 20%, MAE of 3.63, MAPE of 5.75%) posed a challenge, indicating that job satisfaction was influenced by complex factors not fully captured by our variables. This led us to pivot toward descriptive analytics, leveraging Power BI to visualize trends and patterns. We designed two key dashboards: the “Turnover Overview” dashboard, and the “Job Satisfaction and Pay Rate” dashboard, to guide management through key insights.

A critical lesson emerged during visualization design: aspect ratios can distort perceptions of trends. Wider graphs presented balanced trends, while narrower ones exaggerated steepness, risking misinterpretation (Peltier et al., 2021). We carefully adjusted our dashboards to balance clarity and accuracy, ensuring axes were scaled appropriately for each salary group to avoid misleading comparisons (Fisher et al., 2021). This analytical journey—from modeling to visualization—enabled us to uncover nuanced insights into turnover and satisfaction, setting the foundation for actionable recommendations.

Key Findings and Insights

Our analytical journey revealed critical insights into VenturaGear’s turnover challenges, combining descriptive analytics with Power BI visualizations. The regression model, despite its limitations, guided our exploration, while carefully designed dashboards provided clear, actionable trends.

1. Turnover Patterns: Only 11% of employees with recorded “EndDates” left VenturaGear; many changed departments, suggesting lower true turnover than initially assumed. The inventory department exhibited the highest relative turnover, while manufacturing showed the lowest. Departures often peaked in the second year, pointing to a mismatch between onboarding expectations and job reality (Lievens et al., 2001).


2. Motivation Dynamics: Lower-level employees rely heavily on extrinsic motivators, such as bi-weekly pay, which boosts short-term satisfaction but fails to foster long-term commitment (Ryan & Deci, 2000; Bénabou & Tirole, 2003).


3. Performance Metrics: Keyboard presses per hour, the only universal performance metric, varies by role and shows no clear link to satisfaction. This metric may create a sense of monitoring, undermining employee autonomy and trust (Ryan & Deci, 2000).


4. Departmental Challenges: The inventory department’s high turnover and low satisfaction highlight unit-specific issues requiring targeted interventions.

These findings emphasize the need for VenturaGear to balance extrinsic and intrinsic motivators to enhance employee engagement and reduce turnover. For a deeper dive into our visualizations, explore our Power BI Workspace.

Actionable Recommendations

Our findings highlight the need for VenturaGear to address both extrinsic and intrinsic motivators to reduce turnover and enhance engagement. Based on our analysis, we propose the following strategies:

Short-Term Actions

• Competitive Wages: Conduct a market analysis to ensure pay rates align with industry standards, as pay frequency and rate are key extrinsic motivators (Ryan & Deci, 2000).


• Increase Pay Frequency: Shift to bi-weekly pay for all employees, if financially feasible, to strengthen the link between work and rewards.

Long-Term Strategies

• Foster Intrinsic Motivation: Provide clear career growth pathways to enhance competence, implement

Conclusion

Our analysis of VenturaGear’s employee turnover highlights the value of data-driven insights for complex workplace issues. By utilizing systematic data preparation and impactful Power BI visualizations, we identified key attrition drivers, including pay frequency and challenges within the inventory department. While our regression model pinpointed important factors, we found that intrinsic motivation plays a crucial role in long-term employee commitment. This process provided VenturaGear with actionable insights and demonstrated the importance of blending academic rigor with practical solutions, paving the way for targeted strategies to foster a more engaged workforce.

References

Arnold, J. (2022). Learning Microsoft Power BI. ” O’Reilly Media, Inc.”.

Bénabou, R., & Tirole, J. (2003). Intrinsic and extrinsic motivation. The Review of Economic Studies, 70(3), 489-520.

Fisher, J., Chang, R., & Wu, E. (2021). Automatic Y-axis Rescaling in Dynamic

Hair, Black, Babin and Anderson (2018), Multivariate Data Analysis (8th edition), Cengage Learning. Print ISBN: 9781473756540

Lievens, F., Decaesteker, C., Coetsier, P., & Geirnaert, J. (2001). Organizational attractiveness for prospective applicants: A person–organisation fit perspective. Applied Psychology, 50(1), 30-51.

Müller, A. C., & Guido, S. (2016). Introduction to machine learning with Python: a guide for data scientists. ” O’Reilly Media, Inc.”.

Peltier, C., McKenna, J., Sinclair, T., Garwood, J., & Vannest, K. (2021). Brief Report:   Ordinate Scaling and Axis Proportions of Single-Case Graphs in Two Prominent EBD Journals From 2010 to 2019. Behavioral Disorders, 47, 134 – 148. https://doi.org/10.1177/0198742920982587

Rasiel, E. M. (1999). The McKinsey Way. McGraw-Hill.

Ryan, R. M., & Deci, E. L. (2000). Intrinsic and extrinsic motivations: Classic definitions and new directions. Contemporary Educational Psychology, 25(1), 54-67.