Asking an Assistant About Your Data, With Checks
Analyse CSV and spreadsheet files with an assistant, and verify every number before you report it.
A taste of a lesson
I uploaded sales data and asked for average order value by region. North looks much higher than expected.
Worth checking before you report it. Ask three things. First: 'How many rows per region did you use?' A tiny sample can swing an average. Second: 'Show the code or steps, and confirm you averaged order values, not regional averages.' Third: 'Are there duplicate order IDs?' Duplicates of large orders inflate results. Then recompute North yourself with a pivot table. If both match, you can trust it more. How many orders does North have compared with the other regions?
Written by the teacher as an example. In your lesson the tutor answers your own questions, and like any AI it can be wrong.
What you will be able to do
- Profile a data file before any analysis and spot problems early
- Ask precise analytical questions and require the steps behind each figure
- Verify key numbers independently with simple spreadsheet checks
- Report findings with method, caveats and what the data cannot show
Lesson plan
- 1 Calculate or estimate? Find out whether your assistant actually computes results. Start
- 2 Profiling the data Understand the file's structure and quality before asking questions. Start
- 3 Asking precise questions Phrase questions so the analysis matches what you need. Start
- 4 Common analytical traps Recognise errors that produce plausible but wrong results. Start
- 5 Independent verification Confirm key numbers outside the assistant. Start
- 6 Honest reporting Present findings with method, limits and reproducible steps. Start
Try asking
About this tutor
Some assistants can load a data file, run calculations and draw charts on request. That is useful, and risky: the assistant may misread a column, parse dates wrongly, average averages or silently drop rows, and the result still looks professional. This tutor is for people who already know spreadsheet basics and want to use an assistant for real analysis at work. You learn to profile a file first, ask questions precisely, require the steps or code behind each figure, verify key numbers independently, and report findings with honest caveats. It also covers anonymising data before upload.
Reviews
Students can review a tutor after a paid lesson. Nobody has yet.
About the teacher
Spreadsheets, slides, planning and privacy: AI assistants for the busy middle of the workday
9 tutors 358 lessons taught Sample
I teach people how to use AI assistants for the structured parts of office work: spreadsheets, presentations, plans, comparisons and short research briefs. I came from finance administration and project coordination, where a wrong number in a slide is a real problem, so my lessons lean towards checking and verifying. I like concrete exercises: a messy table to clean, a...
See Diego's profile and tutorsMore like this
Other tutors on the same or nearby topics.