๐Ÿ†Data Quality Checker

Get a Data Quality Score
for Any CSV in Seconds

Upload a CSV and get a weighted 0โ€“100 score across completeness, uniqueness, type consistency and statistical integrity plus type mismatches and 3-sigma outliers with exact row numbers, an inferred schema, and a downloadable JSON report.

โœ“ CSV files onlyโœ“ Type mismatch + outlier detectionโœ“ Plain-English insightsโœ“ No upload to serverโœ“ Free forever
๐Ÿ“‚

Drop your CSV file here

CSV files only your file never leaves your browser

Supports any CSV regardless of size. No row limit.

How to use

How to Check Data Quality Before Loading into Power BI or a Database

Skipping a quality check is the leading cause of dashboards showing wrong numbers, database imports failing with cryptic errors, and reports that a stakeholder later flags as inconsistent with the source system.

1
๐Ÿ“‚

Upload your CSV

Click or drag your export file onto the upload area. Your file is read locally nothing is sent to any server.

2
๐Ÿ†

Read the score

See your weighted 0โ€“100 score and a plain-English insight card for every issue, with positive confirmations for what is already clean.

3
๐Ÿ”ข

Drill into the details

Type Mismatches and Outliers tabs show exact row numbers for values that will break Power BI measures or skew averages.

4
๐Ÿ“„

Export and fix

Standardize headers, copy the inferred schema, or export the full JSON quality report then load into Power BI or your database.

The formula

What Makes Up Your Data Quality Score

The score starts at 100. Each of the four factors below can deduct points up to its weight, and the final score is max(0, round(100 โˆ’ total deductions)). A score of 90+ is Excellent (ready for production), 70โ€“89 is Good (minor issues to review), and below 70 is Poor (multiple problems that will cause errors downstream).

Completeness

40%

Deduction = (100 โˆ’ overall fill rate) ร— 0.4. A dataset with 80% overall fill rate loses 8 points.

Uniqueness

30%

Deduction = duplicate row % ร— 1.5, capped at 30. Reaches the cap once 20% of rows are duplicates.

Type Consistency

20%

7 points lost for every column where the dominant type (number/date/boolean) has exceptions, capped at 20.

Statistical Integrity

10%

3 points lost for every numeric column containing values more than 3ฯƒ from the mean, capped at 10.

What this tool catches

Type Mismatches and Outliers Issues Other Tools Miss

Duplicate rows and missing values are well-known problems with dedicated tools. These two checks are different they only surface when you analyse the shape of the data within each column, and they are the most common cause of "the data looks fine but the dashboard is wrong."

๐Ÿ”ข

Type Mismatches

This tool first determines the dominant type of each column number, date, boolean or text by checking what percentage of non-empty values match each pattern (numbers handle commas, currency symbols and percent signs; dates match common YYYY-MM-DD and DD/MM/YYYY formats). If a column is at least 70% one type, every value that does not match is flagged as a mismatch with its row number. An "amount" column with 997 numeric rows and 3 rows containing "N/A" or "pending" will have those 3 rows listed exactly. In Power BI, a single non-numeric value forces the entire column to be typed as text, silently breaking every SUM and AVERAGE. In SQL, the import simply fails.

๐Ÿ“Š

Statistical Outliers (3-Sigma Rule)

For every column classified as numeric, the tool calculates the mean and standard deviation of its values (excluding any type-mismatch rows), then flags any value more than three standard deviations from the mean the standard threshold used in statistical process control. A "quantity" column averaging 12 with a standard deviation of 3 would flag a row containing 9,999 as an outlier roughly 3,329ฯƒ away. Outliers are not automatically wrong, but a single typo like an extra zero can shift an average enough to mislead an entire report, and this check surfaces exactly which row to verify.

FAQ

Frequently Asked Questions

What is a data quality score and how is it calculated? +
The score starts at 100 and points are deducted across four weighted factors: completeness (up to 40, based on overall fill rate), uniqueness (up to 30, based on duplicate row percentage), type consistency (up to 20, 7 points per column with mixed types) and statistical integrity (up to 10, 3 points per numeric column with 3ฯƒ outliers). The result is rounded to the nearest whole number with a floor of 0. 90โ€“100 is Excellent, 70โ€“89 is Good, below 70 is Poor.
What does this tool check that the Find Duplicates and Null Value Checker tools do not? +
Type mismatches and statistical outliers neither dedicated tool covers these. This tool also combines duplicate counts, missing-value rates, type mismatches and outliers into one weighted score with plain-English insights, an inferred schema and a downloadable JSON report. For deep duplicate removal or null-filling with multiple fill strategies, this tool links directly to those dedicated tools.
What is a type mismatch and why does it break Power BI and SQL imports? +
A type mismatch is a value that does not match the dominant type of its column most commonly a text value like "N/A" in an otherwise numeric column. SQL rejects imports when a numeric column receives text. Power BI infers the whole column as text the moment it sees one non-numeric value, breaking every SUM, AVERAGE or other numeric measure on that column.
How does the statistical outlier detection work? +
For every dominantly-numeric column, the tool computes the mean and standard deviation, then flags any value where |value โˆ’ mean| / standard deviation > 3 the 3-sigma rule. Each flagged value is shown with its row number and exactly how many standard deviations it sits from the mean.
Can I export a quality report to share with my team? +
Yes. Export Quality Report downloads a JSON file containing the score and its 4-factor breakdown, every column with its inferred type, fill rate, uniqueness flag and example value, the full lists of type mismatches and outliers with row numbers, the duplicate count, and all generated insights useful as a data handover document or a quality baseline for future comparisons.
What does "Standardize Headers" do? +
It rewrites every column name to snake_case: trimming whitespace, replacing spaces and special characters with underscores, collapsing repeats, and lowercasing the result. "Customer Name " becomes "customer_name" and "Order-ID#" becomes "order_id". Any header that would collide with another after standardization gets a numeric suffix. The tool shows exactly which names changed before you download.
Is my CSV file uploaded to a server? +
No. Your file is read directly in your browser using the JavaScript FileReader API and is never transmitted anywhere. The score calculation, type inference and outlier detection all run locally, which makes this safe for confidential business data.
Does this tool work with Excel files? +
This tool is built specifically for CSV files for fast, reliable browser-based parsing. If your data is in Excel, export the sheet as CSV first (File > Save As > CSV), or use the TOOLBeans Data Profiler, which supports XLSX, XLS and REST API sources directly.
Related tools

More Free Data Quality Tools