Marketing Analyst Agent

An AI-powered marketing analyst agent built with Langchain that automates marketing analysis tasks through a command-line interface.

Overview

This project implements an intelligent agent system capable of analyzing marketing data, generating insights, and providing strategic recommendations - effectively replacing or augmenting human marketing analysts with AI capabilities.

Features

• Market trend analysis and forecasting
• Consumer sentiment analysis
• Campaign performance evaluation
• Competitor analysis
• Automated report generation
• Strategy recommendations
• Data visualization capabilities

Prerequisites

• Python 3.9+
• OpenAI API key (optional when using DEBUG_MODE)
• LangSmith account for tracing and evaluation (optional)

Installation

1. Clone the repository:

git clone https://github.com/yourusername/market_analyst_agent.git
cd market_analyst_agent

2. Create and activate a virtual environment:

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

3. Install dependencies:

pip install -r requirements.txt

4. Set up environment variables:

Create a .env file with the following variables:

OPENAI_API_KEY=your_openai_api_key
LANGSMITH_API_KEY=your_langsmith_api_key  # Optional, for tracing
LANGSMITH_PROJECT=your_langsmith_project_name  # Optional, for tracing
DEBUG_MODE=true  # Set to true for development without API keys

Project Structure

market_analyst_agent/
├── docs/                    # Documentation files
├── src/                     # Source code
│   ├── agents/              # Agent definitions
│   ├── chains/              # LangChain chains
│   ├── tools/               # Custom tools
│   ├── models/              # Data models
│   ├── config/              # Configuration
│   └── utils/               # Utility functions
├── tests/                   # Test files
├── .env                     # Environment variables (not versioned)
├── requirements.txt         # Dependencies
└── README.md                # Project documentation

Usage

Running the agent:

For a single query:

python -m src.main --query "Analyze market trends in the mobile gaming industry"

With specific model:

python -m src.main --model gpt-4 --query "What is the consumer sentiment around Product X?"

With term highlighting:

python -m src.main --query "Analyze mobile gaming market trends" --highlight "growth" "revenue" "engagement"

In interactive mode:

python -m src.main

Using the Convenience Script

A shell script is provided for easier usage with automatic virtual environment management:

# Basic usage (interactive mode)
./run_agent.sh

# Query with highlighting
./run_agent.sh -q "Analyze mobile gaming market trends" -h growth revenue engagement

# Full options
./run_agent.sh -q "Your query here" -m gpt-4 -v -h term1 term2

Available options:

• -q, --query: The query to process
• -m, --model: OpenAI model to use
• -v, --verbose: Enable verbose output
• -n, --no-tracing: Disable LangSmith tracing
• -h, --highlight: Terms to highlight (multiple can be specified)
• -i, --interactive: Force interactive mode

Command-line Arguments

• --query: The query to run (if not provided, interactive mode will be used)
• --model: OpenAI model to use (default: specified in .env)
• --no-tracing: Disable LangSmith tracing
• --verbose: Enable verbose output
• --highlight: Terms to highlight in the response (can specify multiple)

Development Mode

For development and testing, you can enable DEBUG_MODE in your .env file:

DEBUG_MODE=true

In DEBUG_MODE:

• API key validation is bypassed
• A mock LLM is used instead of making real API calls
• Predefined responses are returned based on query patterns

This allows for easier development and testing without requiring valid API keys.

Development

Check the docs/ directory for detailed documentation on each component.

Contributing

We welcome contributions to the Marketing Analyst Agent! Please see our Contributing Guidelines for details on how to get started.

For a comprehensive guide on project structure, documentation standards, and development workflow, refer to our Project Maintenance Guide.

Also, please note that this project is released with a Contributor Code of Conduct. By participating in this project you agree to abide by its terms.

License

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