Browse categories
Explore
Fiverr Pro
English
$
USD
I will clean, deduplicate and normalize your messy dataset in excel
Backend Engineer for Payments, Ledgers and Verified Data
I build the part of the software where money cannot be wrong. Since 2019 I have been the backend engineer behind a cross-border payments platform: remittances, wallets, multi-provider payouts and a do...
About this Gig
The dangerous part of cleaning a dataset isn't the mess you can see. It's the 412 near-duplicates you can't.
Two rows for the same customer with the name spelled differently. Three product entries that are one product. Dates in six formats where two of them are ambiguous, so half your timeline is silently wrong.
Send me the file. You get it back clean, plus a change report: rows in, rows out, what merged with what, and every row I touched. Nothing disappears without appearing in that report.
On near-duplicates: I flag them for your review by default instead of merging them automatically. You know your data. Silently merging two rows that looked similar is how a clean dataset becomes a wrong one, and you'd never find out.
What I fix: exact and fuzzy duplicates, inconsistent dates and number formats, mixed currencies, capitalization and spacing, split or merged columns, broken encodings, malformed URLs and codes, missing values, and merging several files into one consistent dataset.
CSV, XLSX, JSON, TSV. No scraping involved - this is your data, processed.
FAQ
Will you delete rows without telling me?
No. Every removed or modified row appears in the change report, and anything ambiguous gets flagged for your decision instead of being altered silently. You can always reconstruct what I did.
Is my data confidential?
Yes. Used only for your order, never shared, deleted 90 days after delivery unless you ask otherwise. If you need it deleted sooner, say so and I'll confirm when it's done.
What counts as a duplicate in my data?
You tell me, or I'll propose a rule and you confirm before I run it. Same email, same company name, same SKU, fuzzy title similarity - each one gives a different answer, and picking wrong quietly destroys rows. I'd rather ask.
My file is bigger than the top package.
Message me the row count and I'll quote it. Large files are usually cheaper per row, not more expensive, and I'd rather price it properly than have you split it into orders that don't fit.
Do you also scrape the data, or only clean what I have?
This gig is for datasets you already have, from wherever you got them. If you need the data collected from public sources in the first place, message me and I'll tell you honestly whether it's feasible before you order anything.
