🔍 Zurvey

AI-powered survey response validation for market research excellence

Open-ended survey responses are critical for market insights but frequently compromised by:

• Gibberish text
• Copy-pasted responses
• Off-topic answers
• AI-generated content

Manual review processes are labor-intensive, error-prone, and delay valuable insights.

Zurvey leverages advanced AI to automatically validate open-ended survey responses, ensuring data integrity while dramatically reducing manual effort.

• Automated Quality Control: Eliminate gibberish, duplicates, and off-topic responses
• AI Detection: Identify machine-generated content
• Time Efficiency: Reduce validation time by up to 90%
• Enhanced Data Integrity: Make decisions based on authentic human feedback
• Actionable Insights: Generate reliable market intelligence faster

Zurvey employs a sophisticated multi-agent system to thoroughly analyze survey responses:

1. Preprocessing Agent: Handles initial cleaning and normalization
2. Quality Filter Agent: Detects empty, short, or disturbing content (Score 0-1)
3. Relevance Agent: Evaluates response relevance to survey topics (Score 0-5)
4. Complexity & Coherence Agent: Analyzes linguistic structure and sophistication (Score 0-5)
5. Sentiment & Toxicity Agent: Evaluates emotional content and potentially harmful language (Score -2 to 2)
6. AI Detection Agent: Identifies signs of AI-generated responses (Score 0-3)
7. Final Decision Agent: Aggregates all agent scores into a comprehensive quality rating

Each specialized agent applies specific scoring criteria, providing a nuanced evaluation across multiple dimensions of quality. The system then produces a total score to flag responses as either high or low quality based on configurable thresholds.

• Python 3.8+
• Git

1. Clone the repository

git clone https://github.com/hiteshhhh007/Zurvey-Team-Recursion.git
cd zurvey

2. Create and activate a virtual environment (recommended)

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install required packages

pip install -r requirements.txt

Edit the file paths in:

• Codes/Pre-Processing.py
• Codes/Agent-Orchestration.py

python Codes/Agent-Orchestration.py

Configure your quality thresholds in:

python utils/thresholding.py

Generate comparison reports and visualizations:

python utils/comparison.py
python utils/plots.py

1. Pre-processing: Clean and normalize survey response data
2. Multi-agent Validation: Specialized AI agents analyze different quality aspects
3. Scoring: Responses receive quality scores across multiple dimensions
4. Thresholding: Configurable quality thresholds filter responses
5. Reporting: Comprehensive analytics on validation results

Zurvey is proudly developed by Team Recursion, a group of students passionate about improving survey data quality.

This project is licensed under the MIT License - see the LICENSE file for details.

Contributions are welcome! Please feel free to submit a Pull Request.