How is this different from CSV Analyzer?

CSV Analyzer reports each column's inferred type and its unique-value count as a quick text summary. Data Profiler goes further per column: it adds min/max/mean for numeric columns and the most common value with its frequency for string columns, the kind of deeper per-column stats CSV Analyzer doesn't compute.

How is a column's type decided?

Every non-empty value in the column is checked: if all of them are exactly "true" or "false" (case-insensitive), the column is boolean; if all of them parse as a finite number, it's number; if only some of them parse as a number, it's mixed; otherwise it's string. A column with no non-empty values at all is also reported as mixed, since there's nothing to type-check.

Why don't I see min/max/mean for a string column?

Those statistics only make sense for numeric data, so they're only computed for columns typed as number. String columns instead show the most common value and how many times it appears, which is the more useful summary for text data.

Does the empty count include blank cells caused by ragged rows?

Yes. Rows shorter than the header are padded with empty cells before profiling, so a row missing a trailing value is counted as empty for that column exactly the same as an explicitly blank cell.

Data Profiler - Per-Column CSV Statistics

Column	Type	Empty	Unique	Min	Max	Mean	Most common
id	Number	0	5	1	5	3	-
name	String	0	5	-	-	-	Ada (1×)
age	Number	1	5	36	85	54.25	-
active	Boolean	0	2	-	-	-	-
city	String	0	3	-	-	-	London (3×)

Overview

Introduction

Before cleaning, joining, or sharing a CSV, it helps to know more about it than just its column names, how many values are actually missing, whether a column that looks numeric has stray text hiding in it, and what its typical range or dominant value looks like.

This tool profiles every column of a pasted CSV at once, surfacing exactly that: type, emptiness, uniqueness, and type-appropriate summary statistics, without writing a single formula.

What Is Data Profiler?

A per-column CSV profiler that infers each column's type (number, string, boolean, or mixed/empty) from its actual values, then reports how many cells are empty, how many distinct values exist, and additional stats suited to that type.

For numeric columns it reports the minimum, maximum, and mean; for string columns it reports the single most common value and how many times it occurs.

How Data Profiler Works

The CSV is parsed with the shared RFC 4180 grid parser and padded to a uniform rectangle so ragged rows don't throw off column alignment. The first row is treated as the header.

Each column's non-empty values are scanned to infer its type, then min/max/mean is computed for numeric columns (via Math.min/Math.max/an average) or the most frequent value is tallied for string columns.

When To Use Data Profiler

When you've just received a CSV from somewhere else and want a fast sense of its shape and data quality before doing anything else with it.

When a downstream tool or import is failing and you suspect a column has more missing values, or a wider mix of types, than expected.

Often used alongside CSV Analyzer and CSV Size Analyzer.

Features

Advantages

Profiles every column in one pass, no need to write a formula or open a spreadsheet.
Reports type-appropriate statistics, numeric range and mean for numbers, dominant value and frequency for text, rather than one generic stat for every column.
Runs entirely client-side, so the CSV never leaves your browser.

Limitations

Type inference is based purely on scanning the column's own values; it doesn't know a column's intended schema (e.g. a zip-code column of all-digit values will be typed as number).
Only the single most common value is reported for string columns, not a full frequency distribution of every distinct value.

Examples

A mixed dataset with a missing value

Input

id,name,age,active,city
1,Ada,36,true,London
2,Grace,85,true,New York
3,Alan,41,false,London
4,Linus,55,true,Helsinki
5,Margaret,,true,London

Output

age: type=number, empty=1, unique=4, min=36, max=85, mean=54.25. city: type=string, empty=0, unique=3, mostCommon=London (3×). active: type=boolean, empty=0, unique=2.

The age column's one blank cell is excluded from min/max/mean; city's most repeated value (London, 3 times) is surfaced directly.

Best Practices & Notes

Best Practices

Run this before joining two CSVs on a shared column, a surprising "mixed" type on what should be a clean numeric key often points to stray whitespace or a stray text row.
Use the empty count to spot columns with unexpectedly high missing-data rates before relying on that column for analysis.
Re-profile after cleaning a CSV to confirm a fix actually resolved the issue, a column that was previously 'mixed' should read as a clean 'number' or 'string' afterward.

Developer Notes

A column with zero non-empty values is reported as 'mixed' rather than a separate 'empty' type, since the type union only distinguishes number/string/boolean/mixed and there's genuinely nothing to type-check in an all-blank column.

Data Profiler Use Cases

Getting a fast data-quality snapshot of a CSV before further processing
Spotting a numeric column that unexpectedly contains stray text values (reported as 'mixed')
Finding the dominant category in a text column without writing a pivot table

Common Mistakes

Assuming a 'mixed' type means the column is broken; it just means some but not all non-empty values parsed as numbers, worth a look but not automatically an error.
Expecting min/max/mean on a string column, those are only computed for columns typed as number.

Tips

Once you know which columns are clean, feed the same CSV into Dataset README Generator to turn this profile into a shareable schema doc.
A column reported as 'mixed' is worth a closer look with CSV Editor, it usually means a stray non-numeric value snuck into an otherwise numeric column.

Data Profiler