Building Retention Strategies: An Interactive HR Attrition Dashboard For IBM

Skills used: Power bi, Excel

Full Project on Github: https://bit.ly/4aJyvT7

About the data:

Data source for this project is public on Kaggle and can be found at https://www.kaggle.com/datasets/pavansubhasht/ibm-hr-analytics-attrition-dataset

No definite date for the dataset but it was last updated 7 years ago (2017)

The data source is fictionally created by IBM data scientist for practice purposes.

This data source contains 1470 rows of data with 35 columns (No missing values), the data source includes demographic information like age, education, and marital status, as well as work-related attributes like business travel frequency and hourly rate.

This data sourceis interesting as there is potential key insights to be found for HR managers in the age compared to business travel, gender to martial status. Analyzing these potential relationships could help HR managers identify groups at risk of leaving or implement targeted initiatives to improve employee satisfaction.

Question:

Leveraging an IBM HR data source, I designed and implemented an interactive dashboard to empower HR managers with data-driven insights into employee demographics (gender, marital status, education, age) and compensation. This project aimed to identify potential issues with attrition in the company and find causes.

Data Cleaning (Excel):

I used excel to check for any potential issues with the data, such as duplicates, incorrect data and missing values. Some of the techniques used where columns to text, removing duplicates, data validation and find and replace. The data was clean with no duplicates nor any missing values, just 0 values but that is expected in the dataset (Training time last year, Stock options level, Years with current manager, can all have 0 value)

ibm hr dashboard:

This Power Bi dashboard aims to empowers HR managers with comprehensive insights into employee demographics and potential pay gaps. It offers an interactive experience - click on any insight segment (eg. gender, age , education) and every visualization updates, dynamically reflecting the chosen filter. The dashboard features 12 interconnected visualizations.

· Click-to-filter functionality: Explore data intuitively by simply clicking on specific demographics. Every chart, graph, and visual updates, displaying only relevant information for the chosen segment.

·  Interconnected analysis: No need to navigate through multiple menus or drill down levels. All visualizations are dynamically linked, providing a holistic view of the impact of each filter selection.

· Real-time insights: Instantly see how changes in demographics affect salary distributions, department composition, and potential pay gaps, empowering informed decision-making.

· Visually appealing: Light blue color scheme (hex code #118DFF) aligns with IBM branding and promotes readability. Clear labeling and intuitive layout further enhance user experience.

· Visually appealing: light grey used for the worksheet provides a reprieve for your eyes and helps highlight the insights.

· Focus on actionability: Insights are presented in a clear and concise manner, with easy to read labels and text, enabling HR managers to quickly identify potential areas of concern and take data-driven action.

Conclusion:

Our analysis reveals a potential concerning 16.12% employee attrition rate, exceeding industry benchmarks like the recommended 10% annual turnover.  This translates to 237 departures, potentially resulting in lost knowledge, decreased productivity, and higher recruitment costs.

Key Findings:

· Sales representatives exhibit a concerning attrition rate of 39.75%, with 33 individuals leaving the company. Further investigation reveals the average monthly income is $2000, with only 4 sales representatives exceeding $3000. Considering the prevalent commission-based structure of sales roles, this disparity might incentivize departures for more stable income or enhanced compensation elsewhere. Notably, 23 of the departing sales representatives are 30 or under, suggesting potential influence of early career aspirations on their decision.

· 68% (162 individuals) of departing employees had a tenure of 5 years or less, signifying a concerning trend of early-career departures with an additional 23% (54 individuals) left after 6-10 years with the company, demonstrating a second plateau of potential talent loss.

· 45 individuals departing within the 2-5 year range originated from the Research & Development department, representing a significant portion (51%) of this group. A concerning trend lies in the demographic with 43 individuals receiving a monthly salary of $4000 or less, and 25 individuals earning $2000.

· A concerning 59% (79 individuals) of departing Research and Development employees left within the first 3 years of working under their current manager. This is further amplified by 36% (48 individuals) leaving within their first year of working under a new manager, 43 of those individuals are new hires.

Improvements for future iterations of the project, include working with non-generative datasets, as the data is too perfect with no outliers. Improve visuals used to create a more diverse dashboard for more visualizing appealing insights and in-depth analysis.

Full Project on Github: https://bit.ly/4aJyvT7

References:

Apollo technical : https://bit.ly/4bN0P8G

Office Vibe Linkedin: https://bit.ly/4bGdcmF

IBM logo: https://en.m.wikipedia.org/wiki/File:IBM_logo.svg