Excel Random Data Generator
Generate random test data for Excel. Create sample datasets with names, emails, numbers, and more.
Configure Columns
How to Generate Random Data
- Configure columns - name each column and pick its type; number columns take a minimum and maximum.
- Set row count - anywhere from 10 to 10,000 rows.
- Preview - the first ten rows are shown so you can check the shape before committing.
- Download - the workbook you get is the data you previewed, saved as random-data.xlsx with a single sheet called Generated Data.
What Each Data Type Produces
| Type | Value | Cell type in Excel |
|---|---|---|
| Number | A whole number drawn uniformly between your minimum and maximum, both inclusive (defaults 1 and 1000) | Number |
| Text | The word Text followed by a number up to 9999, for example Text 4821 | Text |
| Full Name | One of 20 first names with one of 20 last names - 400 possible combinations | Text |
| first.last plus a number, at gmail.com, yahoo.com, outlook.com, company.com or example.org | Text | |
| Phone | US format (NNN) NNN-NNNN with random area code, prefix and line number | Text |
| Date | An ISO date string between 2020-01-01 and today | Text, not a date serial |
| Yes/No | An even split of true and false | Boolean (TRUE / FALSE) |
Edge Cases and Limits
- No seed, no repeatability. Values come from the browser's Math.random, which is not cryptographic and cannot be replayed. Keep the file if you need the same data twice.
- No uniqueness. Every cell is drawn independently, so numbers, names and emails all repeat - a "random ID" column is not a key.
- Duplicate column names collapse. The column name is used as the field name, so two columns called Amount produce one Amount column in the download. Give each column a distinct name.
- Dates are text in ISO form, so they sort correctly as strings but need converting before date arithmetic works.
- The download reuses the last preview. If you change the columns or the row count after previewing, click Preview Data again before downloading, or the file will hold the earlier set.
- Row count is clamped to 10,000 even if a larger number is typed in.
- Your setup persists in this browser only: the column list and row count are stored locally so the page reopens the way you left it. The data itself is never stored or uploaded.
- Everything is generated in the page, which is why there is no rate limit and no signup - the cost is your own CPU and memory.
Good and Bad Uses for This Data
It is well suited to filling a spreadsheet for a demo, exercising an import routine, testing pivot tables and formulas at a realistic row count, or teaching. It is not suited to statistical work: the distributions are uniform, the columns are independent, and there is no correlation between, say, a name column and an email column in the same row. For a fixed dataset with documented contents and checksums - useful when a test needs to assert on exact values - use the sample Excel files instead.
When to Use a Different Tool
To build a bigger set than one file allows, generate several and combine them with Excel Merge. To turn the output into a fixture for a pipeline, Excel to CSV, Excel to JSON and Excel to JSONL all take the workbook straight to text. To add a formula column across the generated rows, open the file in the Excel editor or build the formula with the formula generator.
Frequently Asked Questions
No. There is no seed: every click calls the browser's Math.random, so each run produces different values. If you need a fixed dataset for a repeatable test, generate it once and keep the .xlsx file - your column setup and row count are remembered in this browser, so the shape is reproducible even when the values are not.
Treat them as strings, not as contact details. Emails are built from four common real domains plus example.org, so an address that looks invented may belong to a real mailbox - never send test mail to them. Phone numbers use the US (NNN) NNN-NNNN pattern with random digits and are not restricted to the reserved 555-0100 to 555-0199 range, so a generated number can be a real line. They are fine for populating a column, filling a form field, or load-testing an import.
The date type writes an ISO string like 2024-07-19, and Excel stores that as text rather than as a date serial. Select the column and use Data > Text to Columns > Finish, or apply a date format after converting, to turn them into real dates. The dates themselves are spread uniformly between 1 January 2020 and today.
Nothing is deduplicated. A number column draws independently from its range each row, so a 1-10000 ID column over 1,000 rows will repeat several values, and the name pool is 20 first names by 20 last names, which repeats quickly. For a unique key, add a column in Excel afterwards and fill a sequence, or remove duplicates with our Remove Duplicates tool.
Between 10 and 10,000 rows per file; a larger number typed into the box is clamped. For bigger datasets, generate several files and combine them with Excel Merge - each file is a separate draw, so the combined set is still random.
Not on this page. The seven types are number, text, full name, email, phone, date and yes/no, with a fixed English-language name list and US phone formatting. There is no locale setting, no address or company generator, and no custom value list - for those, generate the columns you can here and paste your own list into the workbook afterwards.