How to use
- Drop an
.xlsx,.csvor.tsvfile, or paste cells copied from Excel. This page opens with the Column stats & validation step already added. - Once the file is open you get a statistics table with one row per column. The table itself is not changed.
- To check the values, turn on validation rules per column. Rows that break a rule are listed.
- If you need a file rather than a report, stack more steps to clean the table and download that.
What is calculated per column
| Item | Meaning |
|---|---|
| Type | Detected as number, date, TRUE/FALSE, text or empty |
| Count / blanks | Cells with a value, and empty cells (whitespace-only cells count as empty) |
| Unique | Number of distinct values |
| Top values | The 3 most frequent values with their counts |
| Number columns | Sum, mean, median, min, max, standard deviation |
| Date columns | Earliest and latest date |
| Text columns | Shortest and longest length in characters |
A column is treated as numeric (or date) only if every non-empty value looks like one. A single x among the numbers makes it a text column with no numeric stats, which is itself a useful hint that something is off.
Calculation examples
For a column of 1, 2, 3, 4: sum 10, mean 2.5, median 2.5 (an even count, so the average of the middle two), min 1, max 4 and a standard deviation of about 1.2910. It is the sample standard deviation (n-1), the same as Excel’s STDEV.S and not STDEV.P. With a single value no standard deviation is shown.
For 1, 2, 2, 5: 3 unique values, top value 2 (twice), median 2 and standard deviation about 1.7321. A date column holding 2026-10-07, 2026-01-02, an empty cell and 2026-03-04 is reported as a date column with count 3, 1 blank, min 2026-01-02 and max 2026-10-07.
Validation rules
Each column can have these rules:
| Rule | Violated by |
|---|---|
| Required | An empty cell (whitespace-only counts) |
| Type | A value that is not a number, date or TRUE/FALSE |
| Pattern | A value not shaped like an email, phone number or digits-only |
| Unique | A value that appears more than once, and every row holding it is listed |
For example, an email column a@b.com, bad, a@b.com, (empty) with required + email pattern + unique turned on gives 4 violations: row 1 a@b.com (unique), row 2 bad (pattern), row 3 a@b.com (unique) and row 4 empty (required). Row numbers count data rows, not the header.
- The list holds up to 1,000 violations; the total is reported separately.
- The email pattern only checks the
name@domain.tldshape; it cannot tell whether the address exists. - The phone pattern accepts Korean number shapes such as
010-1234-5678,01012345678and+82-10-1234-5678.
Tips
- Right after receiving a file, look for columns with many blanks and for columns whose unique count equals the row count (ID candidates).
- If a column that should be numeric shows up as text, suspect stray values such as
N/A,-or thousands separators. - Comparing stats before and after cleanup shows whether a dedupe or blank-row removal did what you intended.
Related tools
- Several steps at once: Excel & CSV Workbench
- Removing duplicates: Remove Duplicate Rows
- Sorting and filtering: Sort & Filter Spreadsheet
- JSON and CSV conversion: JSON CSV Converter
Your file is processed only in your browser and is never sent to a server.