Data Normalizer
Scale a numeric CSV column with min-max, z-score, or standardization.
About this tool
Pick a numeric column and a scaling mode and get the CSV back with that column rescaled. Min-max maps the values to 0-1. Z-score (or standardize) centres them on 0 with a standard deviation of 1. The mean and SD, or min and max, are returned so you can apply the same scaling later.
How to use
How it works
Min-max: x′ = (x − min) / (max − min). Z-score: x′ = (x − mean) / sd with the sample SD (n − 1). The z-score and standardize options give the same numbers. All other columns are left untouched, and the output includes the parameters used.
min-max: x′ = (x − min)/(max − min) z-score: x′ = (x − mean)/sd
Worked example
Min-max on y
- Input
- Column y: 2.1, 3.9, 6.2, 7.8, 10.1, 12.2, 30 (mode min-max)
- Output
- params: min 2.1, max 30 y: 0, 0.0645, 0.1470, 0.2043, 0.2867, 0.3620, 1
- Walkthrough
- The smallest value becomes 0 and the largest 1. For example (6.2 − 2.1) / (30 − 2.1) = 0.1470.
Limitations
- One column per run, and it must be numeric.
- Missing cells stay missing.
- The scaler is fitted to this data only. It does not separate train and test sets.
- The SD uses n − 1, so values differ slightly from scikit-learn's StandardScaler, which divides by n.
Questions
How do I normalize a column in a CSV?
What is the difference between normalization and standardization?
Will outliers affect min-max scaling?
Can I scale several columns at once?
Features
- Min-max or z-score
- Parameters returned
- Scaled CSV output
Related resources
Guides and videos from around the web—opens on YouTube or the publisher’s site.
SQL Tutorial - Full Database Course
freeCodeCamp.org
SQL and tabular data sit next to CSV/JSON data tools.
Watch on YouTubeFinding mean, median, and mode
Khan Academy
Descriptive stats for quick data summaries.
Watch on YouTube
- Introducing JSON
json.org
JSON grammar for data format converters.
- Model for Tabular Data and Metadata on the Web
W3C
Tabular data model behind CSV-oriented workflows.