Overview

Surveys collect large amounts of data, and manual analysis is time-consuming and prone to missing key information. Claude can help you quickly analyze survey results, calculate statistical metrics, identify trends and patterns, and generate visualization reports.

Use Cases

User satisfaction survey analysis
Market research results organization
Employee feedback statistics
Product requirement prioritization analysis

Steps

Step 1: Data Quality Check

First, understand the basic status of the survey data.

Please analyze ~/surveys/user_feedback.csv:
- Total responses and completion rate
- Number of valid responses for each question
- Identify question types (single choice, multiple choice, text, rating, etc.)
- Check for outliers or invalid responses
- Generate data overview

Step 2: Single Choice Statistics

Analyze the distribution of options for choice questions.

For all single choice questions, please calculate:
- Number and percentage of people selecting each option
- Most popular and least popular options
- Output in both table and text description format
For example: "Product satisfaction" question:
- Very satisfied: 45 people (45%)
- Satisfied: 30 people (30%)
- Average: 15 people (15%)

Step 3: Cross-tabulation Analysis

Analyze response differences across different groups.

Please perform cross-tabulation analysis:
- Group by "Age group" and compare "Purchase intention" differences
- Group by "Usage frequency" and compare "Satisfaction rating"
- Identify groups with significant differences
- Generate comparison tables and key findings

Step 4: Text Feedback Analysis

Analyze text responses from open-ended questions.

For the open question "Your suggestions and comments":
- Extract high-frequency keywords (excluding stop words)
- Identify common themes (such as price, features, service, etc.)
- Categorize feedback by theme, list representative opinions for each category
- Calculate the ratio of positive, negative, and neutral feedback
- Summarize the most frequently mentioned issues and needs

Step 5: Generate Analysis Report

Create a complete analysis report.

Please generate a Markdown format survey analysis report:

# User Feedback Survey Analysis Report

## Basic Information
- Number of collected surveys, completion rate, survey period

## Key Findings
- Top 3 most important findings
- Overall user satisfaction rating
- Main issues and improvement suggestions

## Detailed Statistics
- Statistical charts and analysis for each question
- Cross-tabulation analysis results

## User Suggestions Summary
- User opinions categorized by theme

## Action Recommendations
- Specific improvement suggestions based on data

Save as ~/surveys/analysis_report.md

Tips

When analyzing text feedback, Claude extracts commonalities and trends but cannot completely replace manually reading each response. Important individual cases still require human attention.

If the survey has a "rating matrix" (multiple items using the same rating scale), you can ask Claude to calculate the average score for each item and rank them to quickly identify strengths and weaknesses.

Common Questions

Q: How to handle multiple choice questions? A: Each option in multiple choice questions is counted independently. The percentage base is the number of respondents, not the total number of options, so it may add up to more than 100%. Claude will handle this correctly.

Q: Can time trend analysis be done? A: Yes. If the survey includes submission time, Claude can calculate trends by day, week, or month, and compare differences across different periods.

Q: How to identify fake responses or invalid surveys? A: You can ask Claude to check for: very short response times, same option selected for all questions, obvious duplicate submissions, and other abnormal patterns, and flag suspicious surveys.