Synthetic BLS OEWS Compensation Dataset with Labeled Pay Outliers (FL/AL/GA subset)

Published: 28 October 2026| Version 1 | DOI: 10.17632/rrr2p47s8g.1
Contributors:
Juan Falck,

Description

Synthetic individual-level compensation records (596,856 employees) generated from U.S. Bureau of Labor Statistics OEWS May 2023 wage percentiles for Florida, Alabama and Georgia, with five rule-based types of labeled pay outliers. Each record has a salary and seven one-hot encoded attributes (MSA, state, major and minor occupation group, industry sub-sector and sector, performance rating). The dataset accompanies the paper “CALDM: A Context-Aware Latent Diffusion Architecture for Pay Anomaly Detection in Tabular HR Data.” See README.txt for full documentation.

Files

Categories

Labor Economics, Machine Learning

Licence