Airbnb Analyst Agent 🏠

An intelligent agent that analyzes Airbnb listing data stored in a local MongoDB database. It combines statistical analysis, data visualization, and LLM-powered insights to act as a virtual real estate analyst.

Features

• 🔍 Market Analysis: Generates executive summaries of market conditions, pricing trends, and inventory.
• 💬 Chat with Data: Interactive QA interface. Ask natural language questions like "Find 3-bedroom apartments with strict cancellation policy" and get precise answers derived from your database.
  • Supports complex logic: Filtering, Sorting, and Limits.
  • Full Schema Support: Queries 24+ fields including house_rules, transit, bed_type.
• 📊 Visualizations: Auto-generates charts for Price Distribution, Property Types, and Room Types.
• 🤖 Multi-LLM Support:
  • Providers: OpenAI, Anthropic, DeepSeek.
  • Model Selection: Granular control (e.g., gpt-4o, claude-3-opus, deepseek-coder).
• 🚀 Scalable:
  • Full Dataset Mode: Processes 5,000+ listings locally for aggregation stats.
  • Smart Querying: Uses MongoDB Aggregation Pipelines to keep LLM context efficient.

Prerequisites

• OS: Linux/macOS/Windows
• Database: MongoDB (running locally on port 27017)
  • Database Name: sample_airbnb
  • Collection: listingsAndReviews
• Python: Version 3.10+ (Conda recommended)

Installation

1. Clone/Navigate to directory:

   cd /home/robert/Coding/analysis

2. Create/Activate Conda Environment:

   conda create -n ai python=3.12
   conda activate ai

3. Install Dependencies:

   pip install -r requirements.txt

4. Configure API Keys: Create a .env file in the root directory:

   OPENAI_API_KEY=sk-...
   ANTHROPIC_API_KEY=sk-ant-...
   DEEPSEEK_API_KEY=sk-...

Usage

1. Start the Application:

   python app.py

2. Access UI: Open your browser to http://localhost:7860.

3. Modes:

   • Market Analysis Tab: Select your provider/model, check "Analyze Full Dataset" (optional), and click "Run Analysis" to get a comprehensive report and charts.
   • Chat with Data Tab: Type questions like "What is the most expensive listing?" or "Show me apartments in Porto under $100".

Project Structure

• analyst.py: Core logic class. Handles DB connections, data fetching, statistical calculations, plot generation, and LLM interaction.
• app.py: Gradio web interface. Handles UI state, model dropdowns, and user interaction.
• .env: Stores sensitive API keys (not committed to git).
• requirements.txt: Python package dependencies.
• static/plots/: Generated chart images.

Technologies

• Backend: Python, PyMongo, Pandas
• Frontend: Gradio
• AI Integration: OpenAI SDK, Anthropic SDK
• Visualization: Matplotlib, Seaborn