Prompts for analyzing data
Prompts for data exploration, chart selection, and A/B test interpretation.
Most data work stalls in the gap between having numbers and knowing what they mean. AI is genuinely useful here because it can hold the tedious parts in working memory while you stay focused on the decision: it will draft a vectorized pandas snippet for a fiddly transformation, talk you out of the wrong chart, or walk a budget-versus-actual variance back to a handful of root causes worth chasing.
The prompts in this collection lean into that. Some are calculators with judgment attached — unit economics like CAC, LTV, and payback, or the margin and break-even math behind a price change. Others are about communication: choosing the right chart for your message, writing a metric definition nobody can argue with, or explaining a statistical concept to stakeholders who don't want the jargon.
The pitfall to watch is misplaced trust. A model that interprets an A/B test will happily sound certain about a result that's underpowered or contaminated by a peeking problem. Treat the analysis as a sharp first reading, not a verdict — re-run the numbers it computes, and make it show the assumptions it leaned on so you can check them.
What makes a good prompt for analyzing data
A strong data prompt gives the model the actual shape of your data and the decision riding on it, not just "analyze this." Tell it the columns and their meaning, the grain (one row per what?), the audience, and the question you're trying to answer — "should we ship this variant?" pulls a far better answer than "interpret these numbers."
The best ones also build in skepticism. Ask the model to flag validity risks, state what would change its conclusion, and separate what the data shows from what it's inferring. For anything quantitative, have it surface the formula and intermediate steps so you can verify the math rather than trusting a confident-sounding total.
Get sharper results
- 01Paste a representative sample of rows (including the messy ones) so the model sees real types, nulls, and edge cases instead of guessing your schema.
- 02For A/B tests, give it sample sizes and the metric definition up front — it can't flag an underpowered or peeking-biased result it never saw.
- 03When it computes anything, ask it to show the formula and intermediate numbers so you can re-derive the total yourself.
- 04Before charting, state the one sentence you want a viewer to walk away with; the right chart follows from the message, not the other way around.
Common questions
Can I trust the calculations AI gives me?
Treat them as a fast first pass, not a final answer. Language models can make arithmetic and aggregation mistakes, especially over many rows. Ask for the formula and intermediate steps, then verify the final figures in a spreadsheet or with code before anyone acts on them.
Should I upload my whole dataset?
Usually no. A representative sample of rows plus a clear description of every column is often enough for the model to write transformation code or recommend an approach. Avoid pasting sensitive or proprietary data into tools that may train on your input.
How do I stop AI from overstating A/B test results?
Give it the sample sizes, the exact metric, and how long the test ran, then explicitly ask it to flag validity risks and avoid overclaiming. A good prompt produces a ship-or-keep-running call with caveats rather than a confident win declaration.
Mine Voice-of-Customer Language From Reviews
Extract repeatable customer language from reviews — value phrases, pain points, and emotional words.
Explain Budget vs. Actual Variances for the Period
Computes and prioritizes material budget-versus-actual variances and proposes plausible root causes to investigate.
Break Down Unit Economics: CAC, LTV, and Payback
Calculates and interprets CAC, LTV, LTV:CAC, and payback period from acquisition and retention inputs.
Analyze Pricing Changes and Their Margin Impact
Quantifies the margin and break-even impact of a price change and stress-tests it against several volume scenarios.
Stress-Test the Assumptions Behind a Financial Model
Critically reviews a financial model's assumptions for aggressiveness, sensitivity, and internal consistency.
Design a Finance KPI Dashboard for Leadership
Specifies a grouped, audience-appropriate finance KPI dashboard with exact metric definitions and a layout plan.
Write a Pandas Snippet for a Specific Data Transformation
Generates a vectorized, commented pandas snippet for a precise transformation on your actual DataFrame.
Choose the Right Chart Type for Your Data
Recommends the best and runner-up chart type for your message, variables, and audience with encoding guidance.
Interpret an A/B Test Result Without Overclaiming
Interprets A/B test numbers honestly, flags validity risks, and gives a ship or keep-running recommendation.
Write a Precise, Unambiguous Metric Definition
Produces a precise metric definition with formula, grain, inclusion rules, and resolved edge cases.
Build a Data-Cleaning Checklist for a New Dataset
Generates a prioritized, column-specific data-cleaning checklist tailored to your dataset and intended analysis.
Write a Dashboard Spec Before You Build It
Drafts a build-ready dashboard spec with KPIs, layout, filters, drill-downs, and out-of-scope guardrails.
Explain a Statistical Concept to Non-Technical Stakeholders
Translates a statistical concept into a plain-language explanation tied to a real decision for non-technical stakeholders.