How to Use Claude Code for Data Analysis Without Writing Any Code

You open the spreadsheet, stare at 800 rows of sales data, and know something is wrong. Maybe one product is underperforming. Maybe a particular month tanked. But the moment you start trying to write formulas or build pivot tables, you lose an hour and still do not have a clear answer.

Most business owners are in this exact position. They have the data. They just have no fast way to get insights out of it.

Claude Code changes that. You can hand it a CSV file, ask a question in plain English, and get back an actual analysis: trends, summaries, comparisons, even charts. No formulas. No pivot tables. No data analyst. This guide shows you exactly how.

What "Claude Code Data Analysis" Actually Means

There is some confusion worth clearing up. You can use Claude for data analysis in two distinct ways, and both are covered here.

Option 1: Claude.ai web or desktop app (easiest) Upload your file directly to a Claude chat. Claude reads it and you ask questions. This works for most business analysis tasks: summarizing data, spotting trends, answering specific questions about your numbers.

Option 2: Claude Code (more powerful) Claude Code is Anthropic's agentic coding tool. It runs in your terminal and has full access to your file system. You point it at a data file, describe what you want, and it writes and runs the analysis code for you. You never see the code unless you want to. The output is a clean report.

Option 1 is where most people should start. Option 2 is what you reach for when the analysis is recurring, complex, or involves large files.

What Types of Data Claude Can Analyze

Claude can work with:

• CSV files — The most common format. Export from Excel, Google Sheets, your CRM, Shopify, QuickBooks, or any database.
• Excel (.xlsx) — Upload directly. Claude reads the sheet structure and data.
• Google Sheets — Copy the data into a CSV or paste it directly into the chat.
• JSON — Useful for API exports and app data.
• Plain text tables — Paste tabular data directly into the message and Claude will parse it.

For very large files (tens of thousands of rows), Claude Code with the Files API handles them more reliably than the web upload.

Method 1: Analyzing Data in Claude.ai (No Setup Required)

This is the fastest path. All you need is a Claude.ai account.

Step 1: Export your data as a CSV

From Excel or Google Sheets: File → Download → CSV. From Shopify, QuickBooks, or your CRM: use their export function.

Step 2: Upload the file to Claude

Open Claude.ai in your browser. Click the attachment icon (paperclip) and upload your CSV. You can also paste data directly if the table is small.

Step 3: Write your analysis prompt

Do not just say "analyze this." Be specific about what you want to know. Here is a prompt structure that works well:

I have uploaded a CSV file containing [brief description of what the data is]. The columns are [column names]. Please:

1. Give me a summary of what the data shows overall
2. Identify the top 3 trends or patterns
3. Flag anything unusual or that needs attention
4. Tell me the top 3 actions I should take based on this data

The more context you give Claude about your business and what the data represents, the better the analysis. A column called "rev" is less useful to Claude than one called "monthly_revenue_usd."

Step 4: Ask follow-up questions

Once Claude gives you an initial analysis, ask follow-ups:

• "Which product had the highest growth month over month?"
• "Compare Q1 vs Q2 revenue by channel"
• "What was the average order value for customers who bought more than twice?"

Claude holds the data in context throughout the conversation, so you can dig deeper without re-uploading anything.

Method 2: Using Claude Code for Business Data Analysis

Claude Code is more powerful for recurring analysis, large files, and when you want to save the workflow so it can be run again next month.

A YouTube tutorial from Teacher's Tech documented exactly this workflow. Two spreadsheets, one prompt, and Claude Code produced a full business analysis in about 10 minutes, including trend detection, anomaly flagging, and three specific recommendations. No coding was needed from the user.[^1]

Getting started with Claude Code:

1. Install Claude Code from anthropic.com/claude-code
2. Navigate to your data folder in the terminal
3. Run claude to start a session
4. Drop your data file into the project folder and describe the analysis you want

A simple opening prompt for Claude Code:

I want to build a reusable data analysis workflow. When I point you at a data file, check the data for any issues, analyze it thoroughly, find trends and patterns, flag anything unusual, and produce a clear report with specific recommendations. The report should include actual numbers, formatted tables, and the top three things the business should do next.

This prompt, used once, becomes a reusable workflow. The next time you have new data, you just point Claude Code at the new file and it runs the same process.

The 5 Most Useful Analysis Types for Business Owners

1. Sales Performance Analysis

Hand Claude your monthly sales data. Ask it to break down revenue by product, channel, or region. Ask it to compare this period to the previous one and flag what changed.

Upload your sales CSV, then prompt: "Analyze this sales data. Compare month-over-month performance for each product. Which products are growing, which are declining, and what is the biggest driver of revenue change?"

2. Customer Behavior Analysis

Export your customer list with purchase history. Ask Claude to identify your highest-value customers, how often people are buying, what percentage buy more than once, and which acquisition sources produce the best long-term customers.

"Analyze this customer data. Segment customers by total spend. What does the top 20% look like versus the bottom 20%? What acquisition channel produces the highest repeat purchase rate?"

3. Marketing Campaign Performance

Export your ad or email campaign metrics. Ask Claude to compare campaigns by cost per result, identify which audiences or creatives performed best, and tell you where to reallocate budget.

"Here is my campaign performance data for the last 90 days. Which campaigns had the lowest cost per conversion? Which had the highest ROAS? What would you recommend cutting or scaling?"

4. Inventory and Operations Analysis

For product businesses, paste your inventory turnover data. Ask Claude to flag slow-moving stock, identify products at risk of stockout, and suggest reorder priorities.

"Analyze this inventory report. Which products have the lowest turnover rate? Which are likely to stock out in the next 30 days based on current sales velocity?"

5. Survey and Feedback Analysis

This one surprises most people. Paste open-ended survey responses into Claude and ask it to identify common themes, categorize responses, and summarize what customers are saying.

"Here are 200 customer survey responses to the question 'What could we improve?' Identify the top 5 themes. How many responses mention each theme? What is the most urgent issue based on frequency and sentiment?"

Getting Charts and Visualizations

Claude can generate visualizations using Claude Artifacts. When you ask for a chart, Claude opens an Artifacts panel and renders an interactive version directly in the chat.

Prompt examples for charts:

"Create a bar chart showing monthly revenue by product for the last 12 months."

"Build a line chart showing customer growth over time. Use data from this CSV."

"Create a visualization comparing conversion rates across the three marketing channels in this data."

Artifacts produce charts as interactive HTML. You can download them, screenshot them, or embed them in a presentation. This is genuinely useful for reporting to stakeholders or dropping into a Notion or Google Slides deck.

For charts inside Claude Code, it writes and runs visualization code using Python libraries. The output is a saved image file you can open immediately.

A Real Step-by-Step Example

Here is a complete example using a 12-month sales CSV from a small retail business.

The file contains: date, product name, quantity sold, revenue, channel (online vs. in-store), and customer ID.

Step 1: Upload to Claude.ai

Step 2: Opening prompt:

This is 12 months of sales data for a retail business. Columns are: date, product, quantity_sold, revenue, channel (online or store), customer_id. Give me a full summary of performance. What are the top products by revenue? Is there a trend in online vs. in-store? Which months were strongest and weakest? Flag anything unusual.

Claude's analysis covers:

• Total revenue and units sold
• Top 5 products by revenue and by volume
• Month-over-month trend with the strongest and weakest periods identified
• Online vs. in-store split over time (with a chart if requested)
• An anomaly flag: one product had a 60% revenue drop in October with no corresponding volume drop, suggesting a pricing error

Step 3: Follow-up:

"Drill down on that October pricing anomaly. What was the average price per unit for that product across all months? What happened in October specifically?"

Claude isolates the product, calculates the average unit price by month, and confirms the price dropped sharply in October. The business owner catches what was likely a discount applied to the wrong product.

This entire analysis took under 10 minutes and identified something that would have taken hours to find manually.

Tips That Make Claude's Analysis Much Better

Give columns meaningful names before uploading. "rev" tells Claude nothing. "monthly_revenue_usd" tells it everything. Rename columns before exporting the CSV.

Provide business context. "I run an e-commerce store selling skincare products to women aged 25–45 in the UK" gives Claude the context to make relevant recommendations rather than generic observations.

Ask for numbers, not just descriptions. "Summarize the trend" gives you a paragraph. "Tell me the month-over-month percentage change for each product" gives you a table.

Ask Claude to check data quality first. Before diving into analysis, ask: "Are there any missing values, duplicates, or obvious errors in this data?" This saves you from drawing conclusions from dirty data.

Use Claude Code for recurring reports. If you run the same analysis monthly, build the workflow once in Claude Code. Point it at the new file each month. The same analysis runs in minutes without rebuilding the prompt from scratch. For more on setting up reusable workflows, see the guide on how to use Claude Code to build a website — the workflow-building principles are the same.

Honest Limitations

Claude is excellent at analysis, but it has real limits worth knowing:

File size. The web interface works best with files under ~5,000 rows. For larger datasets, use Claude Code with the Files API or split the file before uploading.

Real-time data. Claude cannot connect to live databases or pull data automatically. You export the file, then analyze it. For automated pipelines, the Claude API lets you wire Claude into a workflow that runs on a schedule.

Statistical precision. Claude is excellent at trend identification, summaries, and comparative analysis. For rigorous statistical tests (ANOVA, regression, significance testing), always verify the outputs or run the code in a dedicated tool.

Memory across sessions. If you close the chat and open a new one, you need to re-upload the file. Claude Projects lets you persist the file and context across multiple sessions, which is worth setting up for ongoing analysis work.

What I Use This For

Personally, what I reach for most is the marketing performance analysis. I export a week of campaign data as a CSV, upload it to Claude, and ask it to tell me what to cut and what to scale. What used to take me 45 minutes of spreadsheet work now takes 5 minutes and often surfaces something I would have missed.

The survey analysis is the other one that surprises people. If you have ever tried to manually code 200 open-ended responses, you know how brutal it is. Claude does it in under a minute and the theme identification is genuinely good.

Start With One File

Pick one spreadsheet you already have and run through Method 1 today. Do not overthink the prompt. Upload the file, describe what it contains, and ask Claude what it sees. The first analysis you get back will show you immediately what is possible.

Download the free Data Analysis Prompt Pack below. It includes 25 copy-paste analysis prompts, 15 chart prompts, and a reusable report template you can hand Claude every time you run a new analysis.

[^1]: https://www.youtube.com/watch?v=we4J_JgPeQA