ML & Data Prep Tools

Data Normalizer opens this category because feature scaling is the step machine learning tutorials mention once and then quietly assume is done. Feed raw house prices ($85k–$420k) alongside a binary flag (0 or 1) into a k-nearest-neighbour model without normalising and the price column drowns the flag completely — not because price matters more, but because its numeric range is five orders of magnitude wider. Min-Max Scaling maps every value into a bounded range you choose (default [0, 1]), which is what neural network activation functions and image pipelines expect. Z-Score Standardisation re-centres to zero mean and unit variance, which is what linear regression, PCA and any algorithm that assumes a Gaussian prior actually needs. The tool handles both in a single paste: comma, space, semicolon, pipe, tab or newline delimiters all work, negative values are supported without special handling, and a single outlier in the sensor-reading sample (10, 12, 11, 13, 10, 99, 11, 12) visibly compresses every other value below 0.04 under Min-Max — the exact behaviour that motivates reaching for Z-Score or outlier removal instead.