Dirty String Standardizer — Free Variant-Entry Cleaner

This tool cleans messy text entries so the same thing is spelled the same way every time. It trims extra spaces, collapses double spaces, fixes the case style, and applies your variant map — so NYC, new york, and NEW YORK CITY can all become New York. Messy variants are the quiet killer of counts: a survey that records New York three different ways reports three small cities instead of one big one. Standardizing first makes every later count, filter, and chart honest.

How to use it

Paste the messy entries, one per line.
Write your variant map as variant = Standard, one per line.
Choose a case style: Title Case, lowercase, UPPERCASE, or keep as typed.
Click Standardize and compare the before and after columns.
Copy the clean column back into your dataset.

Worked example

A conference signup sheet lists home cities as “ NEW york city”, “NYC”, “new york”, “Los Angeles ”, and “LA”. Without cleaning, a count by city gives five tiny groups. With the variant map NYC = New York and LA = Los Angeles plus Title Case, the same list becomes New York three times and Los Angeles twice — the true picture. The variant map is the key step: automatic case fixes cannot know that NYC means New York, so you supply that knowledge once and the tool applies it everywhere.

Honest limits

The tool standardizes spelling, not meaning. It will not catch typos like “New Yrok”, will not merge genuinely different places with similar names, and will not resolve abbreviations you did not put in the map. Accents and non-English characters are preserved but case conversion follows simple rules that may not suit every language. Always scan the after column: a wrong map entry applied to hundreds of rows is worse than no cleaning at all, because it looks clean while being wrong.

Who it is for

This tool is for anyone cleaning survey responses, signup sheets, or scraped lists before analysis — researchers, event organizers, and data-entry teams. It is best for short categorical fields like city, department, or product names, where a small variant map covers most of the mess. For free-form sentences, it only tidies spacing and case; meaning still needs a human.

Going further

Build the variant map iteratively, not in one heroic pass. First standardize with an empty map and count the results: the most frequent raw values are your map candidates. Write mappings for the top variants, re-run, and watch the tail shrink. Stop when the remaining variants are rare one-offs that a map entry would not pay for; hand-fix those in the dataset itself. Keep the map as a file next to your data, with a date and a note about who approved ambiguous mappings, because the same messy file will arrive again next semester or next quarter and the map is the reusable part of your work. Audit after every run by sorting the after column and scanning the distinct values: a clean column has a short, sensible list of distinct values, while a dirty one hides behind a long tail of near-identical spellings. If two distinct values differ by one letter and you cannot explain why both should exist, one of them is probably a typo wearing a map entry it does not deserve. Watch out for over-standardization, the mirror-image error. Merging distinct values because they look alike — two clinics with similar names, a product and its successor — destroys information just as surely as leaving variants apart. When a mapping is uncertain, check an independent field before applying it: the same city with different postcodes, or the same name with different departments, is a warning to split, not merge. A conservative rule serves most projects well: standardize only what you can verify, leave the rest visibly messy, and let the mess remind the next analyst exactly where judgment is still needed.

Embed this tool on your website — free

Copy the code below to add this working tool to your own page. A small credit link stays with it — that’s all we ask.