Excel Data Profiler
Inspect spreadsheet quality before analysis with missing-value counts, uniqueness, duplicate values, type hints, and sample values for every column.
Drop your file here or click to upload
Supports .xlsx, .xls, .xlsm, .xlsb · up to 50MB · or paste a file from your clipboard
What the Profile Computes
The Profile tab builds one row per column of the selected sheet and fills in seven fields. Nothing is modified: this is a read-only inspection you run before deciding what to clean.
| Field | Exactly what it is |
|---|---|
| Column | The header text, or Column 1, Column 2 when First row headers is off. Repeated headers are disambiguated as Region, Region 2. |
| Type | number, text, boolean, mixed, or empty - from the stored types of the non-blank cells. |
| Filled | Cells that are not null and not whitespace-only. |
| Missing | Data rows minus Filled. |
| Unique | Distinct filled values, compared trimmed and case-insensitively. |
| Duplicate Values | Filled minus Unique - the number of surplus occurrences. |
| Sample | The first three filled values, comma-separated. |
Reading the Numbers
The three most useful readings are quick. A high Missing count on a column you intended to join or filter on tells you the join will silently lose rows. A Unique count equal to Filled means the column is a usable key - customer ID, order number - while a Unique count of 1 means the column is a constant and carries no information. A mixed type is the loudest warning on the page: SUM and AVERAGE quietly skip the text cells in that column, which is where most wrong totals come from.
Because Unique ignores case and surrounding spaces, it reports the count you actually want on messy data: a column containing Acme, acme, and ACME profiles as 3 filled, 1 unique, 2 duplicate values. That is also the comparison Remove Duplicates uses by default, so the profile predicts how many rows a dedupe would drop.
Edge Cases
Blanks are never counted as a unique value, so a column of 50 rows with 10 empties and 40 identical entries reports 40 filled, 10 missing, 1 unique. Dates are typed as number because Excel stores them as serials; a date column typed as text is the finding, not a bug. Numbers stored as text - the classic import artefact - make an otherwise numeric column read as mixed. Columns are indexed by the widest row on the sheet, so a stray value far to the right creates real, mostly-empty columns in the report, which is itself worth knowing before you send the file on. The profile covers one sheet at a time; switch sheets in the file bar to profile another.
How to Use It
- Upload the workbook - .xlsx, .xls, .xlsm, or .xlsb, read in your browser.
- Confirm the header setting - untick First row headers if the sheet starts straight into data, or the first record is consumed as labels.
- Scan Type and Missing first - they decide whether the sheet is fit to analyze at all.
- Download Profile - the seven-column report saves as its own workbook, handy for attaching to a data-quality ticket.
When a Different Tool Fits Better
The profiler counts problems; it does not fix them. Once you know what is wrong, Excel Data Cleaner trims and fills, Find Duplicates shows you which rows the repeats actually are, and Convert Text to Number repairs a column that came back mixed. If you want the unique count as a live formula in the sheet rather than a one-off report, use Count Unique Values in Excel.
Frequently Asked Questions
No. The Profile tab only counts and reports - it writes nothing back and offers no cleaned copy of your data. The only download it produces is the seven-column report itself.
Distinct non-blank values in that column, compared after trimming surrounding spaces and ignoring letter case. So Acme, acme, and ' ACME ' count as one unique value, and blanks are excluded rather than counted as a value of their own.
Extra rows. It is Filled minus Unique, so a value appearing three times contributes 2. A column of 100 filled cells holding 98 distinct values reports 2, not 2 duplicated values.
Because that is what it is in the file. Excel stores a date as a serial number with a date format on top, and the profiler reads the stored value, not the display format. A date column typed as text means the dates were never converted to real dates - the usual cause of sorting and SUMIFS problems.
The column holds more than one of number, text, and boolean among its non-blank cells - typically a numeric column with a few cells like N/A or 1,200 stored as text. Those cells are skipped by SUM and AVERAGE, sort differently from their neighbours, and are the most common cause of a total that looks wrong.