SMEMinds SMEMinds
Sign in Start learning
Data Beginner 18 min

Cleaning a messy sheet — the two hours before analysis can even start

The problem

You export sales data. 'Sharma Traders', 'Sharma traders', 'SHARMA TRADERS' and 'Sharma Traders ' are four customers as far as your pivot table is concerned. Half your dates are text.

The workflow
1Copy the file
2Sample 40 rows
3Diagnose
4Get the recipe
5Apply to the copy
6Check counts
7Approve merges
★Save the recipe

What you need

Data: monthly sales export from Tally, 2,400 rows. Want to do: sales by customer by month, and top 20 customers.

The tool

Any assistant, with 30–50 sample rows

The prompt

Subscribers. The full prompt is part of the SMEMinds AI Playbook. See plans.

What comes back

1. COLUMN AUDIT Date — should be date; is mixed text/date in three formats (dd/mm/yyyy, d/m/yyyy, dd-mm-yyyy).

Verify before you use it

Human review required. Check every line below before this leaves your screen.
  • Answer the Sharma Traders, Nashik question yourself before any merge. This is a business fact the model cannot know, and getting it wrong corrupts the top-20 list either way.
  • Check the row count before and after every fix. A cleaning step that drops rows is the most common silent failure.
  • After the customer-name fix, count distinct customers before and after. If it fell from 340 to 180, something over-merged.
  • Ask specifically for the list of ambiguous dates in the full file, and resolve them against the source system rather than assuming.