Dirty Jobs

Data cleanup · Spreadsheets · Small automations

We do the dirty data jobs.

Send the spreadsheet nobody wants to touch. I clean it, keep your IDs and ZIP codes exactly as they were, and hand back a list of every cell I changed. If a value is missing, I flag it. I don't make one up.

  • Fixed price agreed before work starts
  • 1–3 days for most jobs
  • Your original is never edited
Sample Dirty Jobs work order showing rows in, rows out, protected columns and the change log summary
Every delivery ships with a work order like this one (sample on made-up data).

Services

Six jobs, priced up front

Every service has three tiers. Starter is a read-only check of your file plus a fixed quote, and its fee comes off the full job if you go ahead.

CSV & Excel cleanup

Duplicates, mixed date formats, stray spaces, names in three spellings. Up to 5,000 rows with a cell-by-cell change log.

Starter $40 · Full job from $105

Excel & Sheets formula fixes

VLOOKUPs that return #N/A, SUMIFS that miss rows, sheets that broke when someone inserted a column. Up to 5 formulas, each one documented.

Starter $40 · Full job from $115

PDF tables to Excel

Text-based PDF tables into a clean sheet, with the source page on every row and totals checked against the PDF.

Starter $40 · Full job from $115

Shopify & CRM catalog cleanup

The same product under three SKUs, variants with no parent, one customer with four records. IDs, prices and stock stay untouched.

Starter $40 · Full job from $125

Zapier, Make & n8n repair

A workflow that used to run and now fails, skips or doubles up. I fix it in your sandbox and hand back the tests and a rollback note.

Starter $40 · Full job from $155

Python file-processing scripts

The weekly export you reshape by hand, turned into one script you can run yourself. Comes with a README, sample files and tests.

Starter $40 · Full job from $145

Bigger files or unusual jobs get their own quote. Prices are in USD and can differ slightly by marketplace because of platform fees.

The checks

Four checks run before you see anything

They catch the mistakes that make people afraid to hand off their data.

  1. 01

    Every change is listed

    Each edited cell shows up in the change log with its row, column, old value and new value.

  2. 02

    Rows out match rows in

    If the counts differ, every missing or merged row is explained by row number.

  3. 03

    Protected columns don't move

    IDs, ZIP codes, SKUs and account numbers you mark as protected come back byte for byte, leading zeros included.

  4. 04

    A second pass re-reads the original

    A separate check compares your file with the output and fails the job if anything changed outside the rules we agreed.

How it works

Three steps

  1. Send a sample

    Column headers, about 50 rows, the row count, and what "clean" means to you.

  2. Get a fixed price

    You get one price, one delivery date, and the rules I'll follow. No hourly meter.

  3. Get the file and the log

    The cleaned file, your untouched original, the change log, and a list of anything flagged for you to decide.

Samples

What a delivery looks like

These samples use made-up data, so no client files are shown. Each one includes the messy input, the cleaned output, the change log and the check results.

Sample work order PDF Work order (PDF) Rows in and out, protected columns, the four checks, and the change log summary.
rowcolumnbeforeafter
1name" aster stone "Aster Stone
1email"ASTER@EXAMPLE.INVALID "aster@example.invalid
2stateorOR
3email(blank)REFUSED: no source value
2, 4emailsame addressFLAGGED: kept, not merged

An excerpt from the CSV cleanup sample's change log. Rows in and out: 4 / 4. Customer IDs and ZIP codes came back unchanged, leading zeros included. The blank email stays blank, and the duplicate address is flagged for the client to decide.

Jesse Aust

Who does the work

Jesse Aust, Washington State

I've spent 20 years in tech, in datacenters, virtualization, development, engineering and data quality. In a datacenter nothing changes without a record and a way back. I run spreadsheets the same way.

Software does the repetitive passes and runs the checks. I read every flagged item and every check result before a file goes back to you.

Limits: I don't log into your live systems. Send an export and you import the result. I don't merge records that only look alike unless you give me the rule.

Get a quote

Hire me where you already work

Every marketplace below holds your payment until the work is delivered. Or send your file details by email and I'll reply with a fixed price, usually within a few hours.

Email contact@dirtyjobs.pro