Data Generation Tool

Data Generation Tool creates synthetic datasets from a user-defined schema to support testing, benchmarking, and privacy-preserving workflows. It targets developers, data scientists, QA engineers, and platform teams needing repeatable, realistic-looking data without exposing real records. Reproducibility is ensured through a user-specified seed, enabling identical outputs across runs.
Core capabilities include schema-driven field definitions (name, type, required/optional); per-field distributions (normal, uniform, lognormal, categorical, fixed); constraints (ranges, lengths, uniqueness); optional null_rate and simple field correlations; seeds for reproducibility; multiple outputs (CSV, JSON, Parquet). Advanced options allow fixed values, temporal sequences, and lightweight referential rules. Required inputs are the schema and sample size; optional enhancements include seed, null_rate, and correlation settings.
Inputs produce structured data that can feed ETL, analytics, ML model testing, and QA checks; outputs include the dataset and metadata describing distributions and constraints.
The tool uses per-field probabilistic sampling, enforcing constraints and optional correlations; numeric fields use standard distributions, categoricals use weighted choices, and strings may be generated from templates or category pools. Nulls respect the configured null_rate, and output formatting matches the chosen format with locale-neutral numeric representation.