About Data Cleaner
This data cleaner is free, needs no sign-up and runs in your browser, so your data is never uploaded. Paste a table or choose a CSV, TSV or Excel file, switch on the cleaning steps you want and watch the summary change. Download the cleaned table as CSV when you are happy with it.
The steps are the ones most data sets need first: trimming and collapsing whitespace, removing duplicate rows, treating NA and null as missing and then dropping or filling them, turning text such as $1,200 into the number 1200, standardising upper, lower or title case, and find and replace with optional regular expressions.
It also finds outliers in a number column with the IQR rule, the z-score or the modified z-score, highlights them in the preview, and can remove those rows. It opens on a deliberately messy sample so you can see every step work.
How to use Data Cleaner
- Open Paste or upload data to load your file, or start with the sample.
- Switch on the steps you need: whitespace, case, types, missing values, duplicates, find and replace.
- To look for outliers, choose a number column and a method, and adjust the threshold if you want.
- Read Before and after for the row counts and what changed, and check the preview below it.
- Copy the cleaned CSV or download it.
Features
- Trim and collapse whitespace.
- Remove duplicate rows, optionally ignoring case.
- Treat NA, N/A, null, nan and a lone dash as missing, then keep, drop, or fill with a value, the mean, the median, the most common value or the value above.
- Fix types: currency and thousands separators, bracket negatives and yes/no, while keeping ids with leading zeros.
- Standardise text to lower, upper or title case.
- Find and replace in one column or all, with whole-cell, case and regular expression options.
- Outlier detection with IQR, z-score or modified z-score, highlighted in the preview, with an option to remove them.
- A before and after summary with counts of every change.
The outlier rules
Each rule turns a column into fences. A value outside the fences is flagged. Quartiles use linear interpolation between sorted values, and the z-score uses the sample standard deviation (n - 1).
IQR: Q1 - k x IQR to Q3 + k x IQR (k = 1.5)
z-score: |x - mean| / s > t (t = 3)
modified z: 0.6745 x |x - median| / MAD > t (t = 3.5)Worked examples
Sample table, trim and remove duplicates ignoring case gives 14 rows become 12
2 duplicate rows were found: "alice smith" repeats " Alice Smith " once spaces and case are ignored, and Bob Khan appears twice exactly.
Sample Salary column, fix types, IQR with k = 1.5 gives 1 outlier: 980000 (fences 39750 to 66950)
"$980,000" becomes 980000 first, then the rule flags it against the other salaries.
Ages in the sample, z-score with threshold 2.5 gives 1 outlier flagged (240)
One age of 240 is almost certainly a typing error, and it also pulls the mean up and the standard deviation up, which is why z-scores are weaker with small samples.
1 to 9 and 100, IQR gives fences -3.5 to 14.5, 100 flagged
Q1 is 3.25 and Q3 is 7.75, so the IQR is 4.5 and the upper fence is 7.75 + 1.5 x 4.5 = 14.5.
Choosing an outlier method
| Method | Typical threshold | Best when |
|---|---|---|
| IQR | 1.5 times the IQR | Any shape of data; not distorted by the outliers themselves |
| z-score | 3 | Data that is roughly bell-shaped and has many values |
| Modified z-score | 3.5 | Small samples or data with a few extreme values |
Good to know
An outlier is not necessarily an error
The rules flag values that are far from the rest. Some are mistakes, such as an age of 240, and some are real, such as one very large order. Look at each before removing it, and note in your report what you removed and why.
What fixing types does and does not do
A column is converted only when at least 90 percent of its filled cells read as numbers, so a text column with one stray amount is left alone. Cells that already are plain numbers are not rewritten, which keeps ids such as 007 intact. Dates are not converted, because the same text means different dates in different countries.
Privacy: how it works
- Your table is read and cleaned by code running in this browser tab. Nothing is uploaded, and your original file is never modified.
Tips
- Trim whitespace first. A trailing space is the usual reason two equal-looking values count as different.
- Turn on "treat NA as missing" before filling, so the placeholders are counted as gaps.
- Filling with the mean pulls values towards the centre and understates spread. The median is safer when outliers exist.
- Download a cleaned copy and keep the original. This tool never edits the source file.
Frequently asked questions
Is it really free?
Yes. There is no price, no trial and no credit card. Ulkaa Tool does the work with your own device, so there is no server bill to pass on to you.
Do I need an account?
No. There is no sign-up and no login. Open the page and use it. An Ulkaa account is not needed for anything on this tool.
Does it run in my browser?
Yes. Your data are processed by code running in your browser tab, on your own device.
Is my data uploaded?
No. Your data are never sent to Ulkaa or to anyone else. You can check this yourself by watching the Network tab in your browser developer tools while you use the tool.
How do I remove duplicates from a CSV?
Load the file, switch on Remove duplicate rows, and optionally Ignore case. The summary shows how many rows were removed, and you can download the result.
How do I fill missing values?
Choose a mode under Missing values: a fixed value, the column mean or median for number columns, the most common value, or the value above. Turn on "treat NA as missing" so placeholders count as gaps.
What is the IQR rule for outliers?
Find the first and third quartiles, take their difference (the IQR) and flag values below Q1 minus 1.5 IQR or above Q3 plus 1.5 IQR.
Which outlier method should I use?
The IQR rule is a good default. The z-score suits large, roughly bell-shaped data. The modified z-score uses the median and is better for small samples with extreme values.
Does it clean Excel files?
Yes, XLSX, XLS and ODS files are read in your browser. The cleaned table is downloaded as CSV, which Excel opens.
Will it change my original file?
No. It reads a copy in your browser and offers a new CSV to download.
Can I use a regular expression in find and replace?
Yes. Switch on Regular expression, and use $1 and so on in the replacement to reuse parts of the match. An invalid pattern simply changes nothing.
Can I convert CSV to JSON here?
No. The file converter at /file/converter/data converts between CSV and JSON. This tool cleans the table.
Popular Data Cleaner pages
- Remove Duplicates From CSVDelete repeated rows and keep the first of each.
- Fill Missing Values in a CSVReplace gaps with a sensible value, and see how many were filled.
- Trim Whitespace in a CSVRemove stray spaces that make equal values look different.
- Change Case in a CSVMake text consistent: Title Case, lower case or UPPER CASE.
- Fix Data Types in a CSVConvert currency text and thousands separators into plain numbers.
- Find and Replace in a CSVReplace values across every column or just one.
- Outlier Detection in a CSVFlag the values that sit far from the rest of a column.
- IQR Outlier CalculatorFind values outside Q1 - 1.5 x IQR and Q3 + 1.5 x IQR.
- Z-Score Outlier DetectionFlag values more than a set number of standard deviations from the mean.
- Modified Z-Score OutliersA robust outlier rule based on the median and the median absolute deviation.