Claude AI: From Spreadsheet Chaos to Clean Data

If you've ever inherited a spreadsheet where dates appear in three formats, names are spelled four ways, and nobody agrees on what half the columns mean, this class is built around that exact file. This is where the work shifts from fixing what's there to building something new. You'll clean a real messy dataset in the Excel add-in, then move to Claude Desktop for the kind of deeper analysis and visualization the add-in isn't designed for. By the end you'll have had the opportunity to practice choosing which tool fits which kind of task, and why.

Learners will have the opportunity to:

1. Open a messy CSV and a messy XLSX and diagnose what needs fixing before touching anything
2. Practice standardizing inconsistent columns, splitting and merging fields, fixing broken date formats, and removing duplicates, with Claude highlighting every cell it changes
3. Work in Claude Desktop to upload a larger CSV file for analysis the add-in isn't designed to handle
4. Watch Claude write and run Python scripts that analyze the data and produce PNG visualizations
5. Build a pivot table conversationally by describing what they want ("enrollment by school and program type")
6. Request charts the same way ("a waterfall chart showing budget variance by department") and work with the result
7. Explore when to let Claude build a chart end-to-end versus when to have it write the formula and build it themselves
8. Practice verifying a cleaned dataset before doing anything further with it
 

Topics

- Diagnosing problems in messy CSV and XLSX files
- Standardizing columns, splitting and merging fields, fixing dates, removing duplicates
- Watching what Claude changes, cell by cell
- Moving to Claude Desktop for larger-scale analysis
- Python-generated PNG visualizations
- Conversational pivot tables and charts
- Telling the add-in and Claude Desktop apart