Data Normalisation Tool

The tool ingests one or more tabular datasets and identifies numeric columns selected for normalization. It supports three scaling methods: min-max normalization to a 0–1 range, z-score standardization to mean 0 and standard deviation 1, and robust scaling based on median and IQR. The outputs preserve original column names and data types while replacing values in the targeted features. Outputs include the transformed dataset and a per-column statistics summary.
Normalization parameters can be applied globally or per-column. For per-column configuration, users map each numeric feature to a method and optional range. When combining datasets, the tool can compute a shared reference distribution or fit independently per source, depending on the workflow. Missing values are handled according to configurable rules and do not break transformations.
Intended users include data engineers, data scientists, analytics teams, and ML engineers who need reproducible data preparation for reporting, dashboards, or model training. The tool differentiates itself with per-column control, auditable parameter export, and batch processing for large data environments, reducing manual scripting and ensuring consistent feature scales across projects.
Common use cases include harmonizing customer datasets from multiple regions, preparing features for regression or distance-based models, and aligning BI datasets for fair comparisons. The tool is designed for speed, transparency, and repeatability, with explicit handling of edge cases and clear error reporting.