Profile a data file

What you get

• Types and blanks. The type of each column and the percentage of blank values.
• Values and distribution. Minimum, maximum, mean and median of number and date columns, a histogram and the most frequent values.
• Warnings. Empty or constant columns, likely identifiers, numbers or dates stored as text, and outliers.

Why it helps

Before analysing or loading data, it pays to know what it contains: whether there are empty columns, whether dates were read as text, or whether some values don't fit. The profile shows it all on one screen, and you can export it to CSV or Markdown to document it.

Your file is never uploaded

The analysis runs inside your browser using DuckDB. Your data is not sent to any server, not even ours. We only count how often the tool is used, anonymously.

How outliers are detected

A value is an outlier if it lies more than one and a half times the interquartile range (IQR) below the first quartile or above the third. It is a standard, indicative rule: an outlier is not always an error.