Data

Spreadsheet Analyst

Turn messy spreadsheets into clean data, clear answers, and honest charts.

by Priya Raghavan

Works with

Google SheetsRequired
Excel
Google Drive

Overview

Hand it the spreadsheet you've been avoiding: the customer list full of duplicates, the sales export with dates in three formats, the survey nobody summarized. It cleans in bulk - dedupes rows, standardizes dates and names, splits and merges columns - and flags what it couldn't fix instead of guessing. Then it does the analysis: totals, trends, comparisons, pivot-style summaries, and charts that make one point each. Everything comes back with a change log, so you can see exactly what was fixed and check the judgment calls before you rely on the numbers.

Try asking

My customer list is full of duplicates and inconsistent names - clean it up and show me what changed
Merge these two contact exports into one list without losing anyone
Here's a year of sales data - what are my trends by month, and which products carry the business?
Turn this survey export into a summary I can put in front of a client
Build a chart of revenue by month that I can drop straight into a slide
Write me a formula that pulls the right price from my price sheet into my quote tab

Examples

This creator hasn't published sample artifacts yet.

Limits & risks

It won't send, spend, publish, or change live systems on your behalf; it only writes to a sheet when you've explicitly attached that connector and told it to
It never deletes rows silently: anything removed or changed is listed in a change log or moved to a review tab for you to check
It won't invent numbers to fill gaps or smooth a trend; missing or contradictory data gets flagged, not guessed
It's not an accountant, auditor, or financial advisor; have a licensed professional confirm tax, payroll, and investment figures before anyone acts on them
It can't recover data that isn't in the file you give it, and it won't scrape websites or buy contact lists to fill the gaps

Getting started

Requirements

Share the data - paste it, upload the file, or connect Google Sheets or Excel - and tell it what a good result looks like for you: which columns matter, what counts as a duplicate (same email? same name and phone?), and the question you want answered. If you don't know, it will ask two or three quick questions and state its assumptions before working.

Steps

1Send the file or connect the sheet, and say what you need: cleanup, a merge, an answer, or a chart
2Answer a couple of quick questions about which columns matter and how it should treat duplicates
3Get back the cleaned data, summary, or chart, plus a change log of every fix it made
4Check the review tab of rows it wasn't sure about and tell it how to handle them
5Keep asking follow-up questions on the same data until you have what you need

What's inside

Model: anthropic/claude-sonnet-4-6

Spreadsheet Analystanthropic/claude-sonnet-5
Hand it the spreadsheet you've been avoiding: the customer list full of duplicates, the sales export with dates in three formats, the survey nobody summarized. It cleans in bulk - dedupes rows, standardizes dates and names, splits and merges columns - and flags what it couldn't fix instead of guessing. Then it does the analysis: totals, trends, comparisons, pivot-style summaries, and charts that make one point each. Everything comes back with a change log, so you can see exactly what was fixed and check the judgment calls before you rely on the numbers.

Spreadsheet Analyst AI agent on Augex