top of page
  • garywalton05

Evaluating Sample Datasets: 'Human Resources Data Set'

The Human Resources Data Set, created by Dr. Rich Huebner and Dr. Carla Patalano, serves as supplemental material for students taking the "HR Metrics & Analytics" course at New England College of Business. While the dataset is synthetic it is not perfect, it includes real-world data issues that can occur in HR environments. For instance, there may be occasional missing manager IDs.

Key Evaluation Criteria:

Data Volume 2/10: The dataset comprises 311 workers, limited to the United States within a single business. It includes basic demographic and job data, along with information on hires, terminations, latest engagement and performance scores, and time & attendance figures. However, the dataset does not provide the wide range of workers and timeframes required for a comprehensive analysis.

Data Completeness 3/10: The dataset contains relevant information for workforce analytics, including dimensions such as demographic and job data. However, it lacks comprehensive fact and event tables, making it challenging to analyse specific scenarios or individual worker histories.

Data Quality 5/10: The dataset is synthetic, however there have been efforts made to inject data issues and inaccuracies. Additionally, as a synthetic dataset of limited scope, it may not reflect real-world scenarios or the intricacies of real-world HR data.

Realism and Complexity 4/10: While the dataset introduces some real-world issues, it is limited to a single business and lacks the diversity of geographies and business units typically encountered in multinational corporations.

Data Availability 10/10: The dataset is publicly available and can be used ethically and legally.


This dataset does not meet the requirements for a comprehensive analysis and portfolio showcasing. Its limitations in terms of worker diversity, limited timeframe, absence of individual worker or job history, and aggregated reporting prevent deeper investigation and analysis.



bottom of page