CSV Profiler

Profile a CSV file and generate a simple summary report per column.

What you get (per column)

• Column name
• Inferred type
• Missing values + non-empty counts
• Top (most frequent) value
• Numeric stats (when applicable): min, max, mean, median
• Unique count

The project includes:

• A CLI built with Typer
• A Gradio web UI

Run without installing (uvx)

If you have Astral's uv installed, you can run the tool without installing it into your environment.

CLI:

uvx git+https://github.com/afaskar/CSVProfiler.git cli data/data.csv

Web UI:

uvx git+https://github.com/afaskar/CSVProfiler.git csv-profiler web

Installation

pip install git+https://github.com/afaskar/CSVProfiler.git

For local development:

pip install -e .

Quickstart (CLI)

csv-profiler cli data/data.csv

This generates the reports in outputs/ by default:

• outputs/report.md
• outputs/report.json

Choose output directory

csv-profiler cli data/data.csv --out-dir output

Choose output format

Supported values for --format are: md, json, both.

csv-profiler cli data/data.csv --format md

csv-profiler cli data/data.csv --format json

csv-profiler cli data/data.csv --format both

Run via Python module

python -m csv_profiler cli data/data.csv

You can also call the CLI app module directly:

python -m csv_profiler.cli cli data/data.csv

Web UI (Gradio)

csv-profiler web

Or via module:

python -m csv_profiler web

Or run the web module directly:

python -m csv_profiler.webgui

Output

• CLI output files are written as report.md and/or report.json inside the output directory.
• Web UI prints the generated report content directly (Markdown or JSON).

License

MIT. See LICENSE.