agentic_trend_comparison

Agentic Trend Comparison: Air Pollution Analysis

This project provides a set of agentic tools to fetch, analyze, and visualize global air quality data. It leverages a Kaggle dataset to provide insights into pollution trends for various cities around the world. The system is built using the FastMCP framework, allowing an agent to interact with the tools programmatically.

✨ Features

• City Discovery: Find cities by name or country, or get a random selection of locations.
• Data Retrieval: Fetch detailed, time-series pollution data for specified cities.
• In-depth Analysis: Calculate overall AQI, identify primary pollutants, and determine long-term trends (improving, worsening, or stable) using linear regression.
• Data Visualization: Generate and save time-series plots to visualize AQI trends over time for one or more cities.

🛠️ Implemented Tools

This project exposes its functionality through a series of MCP tools:

• location_tool: Selects cities based on user input or randomly from the dataset.
• fetch_tool: Retrieves the raw pollution data records for the selected cities.
• analyze_trends_tool: Processes the raw data to generate a summary, including average AQI, primary pollutant, and the pollution trend over time.
• plot_trends_tool: Creates a .png plot of the AQI data over time and saves it to the ~/air_pollution_artefacts directory.

🚀 Getting Started

To run the agentic tool MCP Server, execute the following command from the root directory of the project:

Overview

MCP (Model Context Protocol) Servers allow AI assistants like Claude and Amazon Q developer CLI to access tools and data sources. This repository includes MCP Servers for data fetching and analysis to help you get started with MCP for air data analysis based on city.

To use the Application Modernization MCP Servers, two approaches will be covered:

• Claude Desktop

Prerequisites

1. Clone the current repository:

git clone ....

And navigate to the project folder:

cd pollution_trends

2. Install MCP Python SDK:

--> GitHub Repository

For this:
Install uv package manager:

curl -LsSf https://astral.sh/uv/install.sh | sh

Add project dependencies:

cd application_modernization
uv add "mcp[cli]"

If you encounter an error regarding a missing pyproject.toml when using uv add, you can alternatively install the package using pip:

pip install "mcp[cli]"

This will start the FastMCP server, which listens for tool calls via stdio.

**Setup Instructions for the custom MCP Servers ** Follow these steps to get the custom Code Modernization MCP server up and running.

1. Navigate to Project Folder

cd pollution_trends 

2. Create virtual environment and activate it

uv venv
source .venv/bin/activate

3. Install dependencies

uv add "mcp[cli]" httpx

Using MCP Servers

Option - Claude Desktop

1. Install Claude Desktop application (https://claude.ai/download)
2. Configure Claude to use your MCP server:
   Open the configuration file (Mac): ~/Library/Application Support/Claude/claude_desktop_config.json
   Add your MCP server configuration:

{
  "mcpServers": {
    "code_modernization": {
      "command": "uv",
      "args": [
        "run",
        "--with",
        "mcp[cli]",
        "mcp",
        "run",
        "<full_path_to_code_analysis_modernization/pollution_trends.py>"
      ]
    }
  }
}

Ensure that the paths are correctly specified. Incorrect paths will prevent the servers from starting.

Note: If "command": "uv" doesn't work, use the absolute path to uv instead: /Users/<username>/.local/bin/uv
3. Save json and restart Claude Desktop
4. Your tool should appear in the chat interface
5. Start using your tools and interact with them 🚀

✅ Project Status & To-Do

• Implement Kaggle Dataset integration.
• Develop core data analysis logic.
• Set up the agentic tool structure with FastMCP.
• Implement location_tool, fetch_tool, analyze_trends_tool, and plot_trends_tool.
• Implement a comparison_tool to directly compare the analysis results of multiple cities.
• Filming and documenting demo of how it works.

📊 Data Source

This project uses the World Air Quality Index by City and Coordinates dataset from Kaggle.