Build a Dashboard from Excel Data with Claude
Want to build a dashboard from an Excel workbook with Claude? In this lesson, Tom Fragale demonstrates a workflow using a multi-sheet sample workbook and a detailed prompt. The example includes customer, product and employee rankings, along with charts that make the results easier to review.
The useful lesson is not just asking AI to make something attractive. It is explaining what you want to measure, then checking the generated calculations against the source data before relying on the dashboard.
Watch: Build a dashboard with Claude
Watch the dashboard demonstration on YouTube.
Start with a workbook you understand
A multi-sheet workbook can hold related information in separate places. Before asking Claude to analyze it, identify what each sheet contains and which fields connect the records. Keep an unchanged copy of the source so you can trace any result back to its original rows.
Use sample data for practice. Before uploading real customer or employee information, confirm that sharing it with the AI service is permitted under your organization’s data-handling rules. Remove private information that the task does not need.
Give the dashboard a clear purpose
Tom’s demonstration uses a detailed prompt rather than a vague request to analyze everything. For your own dashboard, specify the questions you want answered and the measures used to rank results. A ranking by revenue can differ from one based on quantity or order count.
Use this companion checklist when writing your request:
- Scope: Name the sheets and date range to include.
- Measures: Define each total, count or other calculation you need.
- Rankings: Explain how to group customers, products or employees, which measure to sort by and how many results to show.
- Charts: Request clear titles, labels and units so readers know what each visual means.
- Uncertainty: Ask Claude to flag missing fields, unmatched records and assumptions instead of silently filling gaps.
Check the calculations before the presentation
A polished chart is not proof that its numbers are correct. Compare the generated dashboard with the source workbook. First, confirm that the intended sheets and records were included. Then independently check totals and a few individual customer, product and employee results.
Look for duplicate records, missing matches between sheets and incorrect date filters. Check that the ranking uses the measure you requested. Finally, compare chart labels and values with the underlying summary. If anything differs, resolve the cause before sharing the dashboard or using it for business decisions.
Treat sample employee rankings as a reporting exercise, not a complete assessment of anyone’s performance. Context and verified data matter. AI-generated summaries should support human review, not replace it.
Practice the full workflow
Start with a small approved sample, write the request, inspect the dashboard and trace several results back to the workbook. Revise the prompt where needed and check the output again. If your source starts as a document, see the related lesson on converting a PDF to Excel with Claude.
Excel and AI training for your team
Want help applying these skills to everyday reporting? Book a free 15-minute private training discussion with Tom Fragale. Explore custom training services, upcoming live webinars and recorded courses.
Contact Tom Fragale
Tom Fragale
Email: tom@pcwebinars.com
LinkedIn: https://www.linkedin.com/in/tomfragale
Website: https://pcwebinars.com