Intellectual Preference Survey

A Python-based survey system that conducts automated intellectual preference surveys using LLM-powered question selection. The system intelligently selects questions based on previous responses, creating a personalized and coherent survey experience.

Overview

This project implements an automated survey system that:

• Conducts surveys using a curated bank of 31 philosophical/theological questions
• Uses Google's Gemini 2.0 Flash model (via OpenRouter API) to intelligently select questions
• Limits each survey to 5 questions, ensuring focused and efficient data collection
• Prevents question repetition within each survey session
• Simulates user responses (yes/no) for testing purposes
• Saves survey results with timestamps in JSON format

Features

• Dynamic Question Selection: Uses LLM to choose the most contextually appropriate next question based on previous responses
• State Management: Tracks conversation history and remaining questions to prevent duplicates
• Persistent Storage: Saves all survey results in a structured JSON format
• Multiple Survey Support: Can run multiple survey sessions in sequence
• API Error Handling: Robust error handling for API interactions
• Configurable: Supports custom API keys and save locations

Prerequisites

• Python 3.7+
• OpenRouter API key
• Required Python packages (see requirements.txt)

Installation

1. Clone the repository:

git clone https://github.com/philipgalebach/intellectual-preference-survey.git
cd intellectual-preference-survey

2. Install dependencies:

pip install -r requirements.txt

3. Create a .env file in the project root with your OpenRouter API key:

OPENROUTER_API_KEY=your_api_key_here

Usage

Basic Usage

Run a single survey:

python survey.py

Run multiple surveys:

python survey.py 5  # Runs 5 survey sessions

Output

Survey results are saved in simulations/survey_results.json with the following structure:

[
  {
    "timestamp": "20250220_234416",
    "responses": [
      {
        "question": "Is doubt part of authentic faith?",
        "answer": "yes"
      },
      // ... more responses
    ]
  }
]

Project Structure

intellectual-preference-survey/
├── survey.py           # Main survey implementation
├── requirements.txt    # Python dependencies
├── .env               # API key configuration
├── simulations/       # Survey results directory
│   └── survey_results.json
└── README.md          # This file

Configuration

The survey system can be configured through:

1. Environment variables in .env:

   • OPENROUTER_API_KEY: Your OpenRouter API key

2. Command line arguments:

   • First argument: Number of survey runs (default: 1)