AI Prompts for Claude for Data Analysis

I need to analyze [DESCRIBE DATASET OR SITUATION] to help [BUSINESS DECISION OR STAKEHOLDER QUESTION]. Help me sharpen the analysis question. What exactly am I trying to measure? What metric best captures it? What would "success" look like in the data, and what would "failure" look like? Also flag any ways my question could be answered superficially without actually solving the underlying business problem.

My analysis question is: [QUESTION]. My data includes: [DESCRIBE COLUMNS AND DATA TYPES]. Suggest an appropriate analysis approach. What type of analysis fits this question: descriptive, diagnostic, predictive, or prescriptive? What statistical methods or visualizations would best answer it? What are the most important assumptions I should check first?

Here is a sample of my dataset (first 20 rows): [PASTE DATA SAMPLE]. And here is a description of what each column represents: [DESCRIBE COLUMNS]. Identify any data quality issues I should address before analyzing: missing values, outliers, inconsistent formatting, likely data entry errors, or columns that may not mean what I think they mean. Suggest how to handle each issue.

My data has these quality issues: [DESCRIBE ISSUES — MISSING VALUES IN X COLUMN, INCONSISTENT DATE FORMATS, OUTLIERS IN Y COLUMN]. I am using [PYTHON/SQL/EXCEL/R]. Write the data cleaning steps I should perform, in order, with code for each step. Explain the reasoning behind each step so I understand what I am doing and can adapt it if my data differs slightly.

I need to analyze [DESCRIBE DATASET AND QUESTION]. I have access to [TOOLS YOU HAVE — PYTHON, SQL, EXCEL, R, TABLEAU, LOOKER]. Which tool is best for this specific analysis? Consider: data size, type of analysis, whether I need to share results, and whether I will need to repeat this analysis regularly. Give me a recommendation with reasoning.

I have a dataset with the following columns: [LIST COLUMNS AND DATA TYPES]. I am using [PYTHON WITH PANDAS / R]. Write a complete exploratory data analysis script that: summarizes each column (data type, missing values, unique values, distribution), identifies correlations between key variables, and produces three to five visualizations that would give me a rapid overview of the dataset. Include comments explaining each step.

I have a database with these tables and relationships: [DESCRIBE SCHEMA]. My business question is: [QUESTION]. Write the SQL query that answers it. If the question requires joins, aggregations, window functions, or subqueries, include those and explain what each part is doing. Also flag any assumptions I need to verify about the data (e.g., that IDs are unique, that dates are consistently formatted).

I am running this code: [PASTE CODE]. It produces this error: [PASTE ERROR MESSAGE]. I am using [PYTHON / R / SQL] and [LIBRARY/DATABASE IF RELEVANT]. Diagnose what is causing the error and show me the corrected code. Explain what the bug was so I understand it, not just so I can copy the fix.

This query or script is running too slowly: [PASTE CODE]. My dataset has approximately [NUMBER] rows and [NUMBER] columns. The query is taking [X] seconds / minutes. Identify the likely bottleneck and rewrite the code to run faster. Explain the optimization technique used so I can apply it to similar situations.

I find myself running this analysis repeatedly with slightly different parameters: [DESCRIBE WHAT YOU DO EACH TIME]. Write a reusable Python function (or SQL stored procedure) that accepts the varying parameters as arguments, runs the analysis, and returns a clean result I can use downstream. Include input validation and a docstring.

I ran a [DESCRIBE STATISTICAL TEST — T-TEST / REGRESSION / CORRELATION / ANOVA / CHI-SQUARE] and got the following result: [PASTE RESULT]. My sample size is [N]. Interpret this result in plain English. Is it statistically significant? Is the effect size meaningful, or is this a significant result that is too small to matter in practice? What are the main limitations on what I can conclude?

My analysis shows that [DESCRIBE FINDING — E.G., CUSTOMERS WHO USE FEATURE X HAVE 40% HIGHER RETENTION]. Before I present this as a causal relationship, help me identify at least three alternative explanations that could explain this correlation without X causing higher retention. How could I design a test or collect additional data to rule out the strongest alternative explanations?

Here is my analysis methodology and findings: [DESCRIBE WHAT YOU DID AND WHAT YOU FOUND]. Check for these common analysis errors: survivorship bias, Simpson's paradox, p-hacking or multiple comparisons problem, confounding variables, selection bias, and overfitting. Flag any that may apply to my analysis and explain how I should address them.

I segmented my data into groups based on [SEGMENTATION CRITERIA] and found the following differences between groups: [DESCRIBE FINDINGS]. Help me validate whether this segmentation is meaningful. Are the groups large enough? Are the differences statistically significant? Is the segmentation criteria likely to reflect a real underlying difference in behavior, or could it be an artifact of how the data was collected?

We use [METRIC NAME] as a key performance indicator, and it currently shows [VALUE]. But I want to stress-test whether this metric is actually measuring what we think it measures. Help me examine: What behaviors or events cause this metric to go up? What could make the metric look good while the underlying reality is getting worse? What data would I need to confirm this metric is a reliable signal of [BUSINESS GOAL]?

My analysis shows the following: [DESCRIBE KEY FINDINGS]. My audience is [DESCRIBE STAKEHOLDERS — EXECUTIVES / PRODUCT TEAM / MARKETING / BOARD]. Write a one-page executive summary that: states the business question we were answering, summarizes the key finding in one sentence, provides the three most important supporting data points, and ends with a clear recommendation or decision that should follow from this analysis.

I need to explain this technical finding to a non-technical audience: [DESCRIBE FINDING]. Write an explanation that uses an analogy, avoids jargon, and focuses on what this means for the business — not how we calculated it. The explanation should be at most three sentences and should answer the implicit question: "So what does this mean for us?"

I have these analysis findings: [LIST KEY FINDINGS]. I need to present them to [AUDIENCE] in [FORMAT — SLIDE DECK / WRITTEN REPORT / FIVE-MINUTE TALK]. Help me build the narrative structure. What is the setup (context and question)? What is the complication (why this matters now)? What is the resolution (what the data shows)? What is the call to action? Outline the presentation in this structure.

I have these findings to visualize: [LIST EACH FINDING]. For each one, recommend the most effective chart type (bar, line, scatter, heatmap, funnel, etc.) and explain why that chart type best communicates this specific insight. If there is a common mistake people make when visualizing this type of data, mention it so I can avoid it.

I am about to present these findings: [DESCRIBE FINDINGS]. My audience includes [DESCRIBE STAKEHOLDERS AND THEIR TYPICAL CONCERNS]. Generate the top seven questions they are likely to ask — including skeptical questions that challenge the methodology or the conclusions. For each question, draft a concise, honest answer that I can prepare in advance.