Market Insights Analyst

An autonomous, multi-agent system that emulates the workflow of a financial analyst. The system ingests and processes a variety of financial data, uses specialized agents to perform research and analysis, and generates a coherent, evidence-based narrative to answer complex user queries, thereby providing actionable market insights.

---

🚀 Features

• Multi-Agent System: Utilizes a collaborative multi-agent framework where each agent has a distinct role, coordinated by a central manager.

• Data Ingestion: Ingests and processes financial news from local files.

• Specialized Agents: Includes a Financial News Researcher Agent and a Stock Data Analyst Agent.

• Evidence-Based Narratives: Generates coherent, evidence-based narratives to answer complex user queries.

• Web Interface: A user-friendly web interface built with Streamlit to interact with the system.

---

🛠️ Technology Stack

Component	Technology
Orchestration	LangChain
Data Framework	LlamaIndex
Vector Store	FAISS (faiss-cpu)
Web Framework	FastAPI
Embedding Model	BAAI/bge-large-en-v1.5
LLM	Gemini 2.5 Pro
Containerization	Docker
Frontend	Streamlit
Evaluation	RAGAS

Export to Sheets

---

⚙️ Getting Started

Prerequisites

• Python 3.8+

• An environment with the required packages installed.

• A Google API key for Gemini.

1. Clone the repository

Bash

git clone https://github.com/your-username/Market-Insights-Analyst.git
cd Market-Insights-Analyst

2. Create a virtual environment and install dependencies

Bash

python -m venv venv
source venv/bin/activate  # On Windows, use `venv\Scripts\activate`
pip install -r requirements.txt

3. Set up your environment variables

Create a .env file in the root directory and add your Google API key:

GOOGLE_API_KEY="YOUR_GOOGLE_API_KEY"

4. Ingest your data

Place your news articles (as .txt files) in the data/news_articles directory. Then, run the ingestion script to create the vector store:

Bash

python ingestion_scripts/ingest_news.py

5. Run the application

The application consists of a FastAPI backend and a Streamlit frontend.

To run the backend:

Bash

uvicorn api.main:app --reload

To run the frontend in a separate terminal:

Bash

streamlit run frontend/app.py

Now, you can open your browser and navigate to the Streamlit URL (usually http://localhost:8501) to start asking financial questions!

---

🚢 Docker Deployment

For a more robust and reproducible setup, you can use Docker. The provided Dockerfile will set up the entire application, including the data ingestion.

1. Build the Docker image:

   From the root of the project, run the following command:

   Bash

   docker build -t market-insights-analyst .
   
   

2. Run the Docker container:

   You'll need to pass your Google API key as an environment variable to the container.

   Bash

   docker run -p 8000:8000 -p 8501:8501 -e GOOGLE_API_KEY="YOUR_GOOGLE_API_KEY" market-insights-analyst
   
   

This command will:

• Start the container.

• Expose port 8000 for the FastAPI backend and port 8501 for the Streamlit frontend.

• The start.sh script inside the container will run both the backend and frontend services.

The Streamlit app will be available at http://localhost:8501.

---

🧪 Evaluation

The project includes a plan for evaluating the RAG pipeline using the RAGAS framework. You can run the evaluation script to measure metrics like Faithfulness, Context Precision, and Answer Relevancy. This helps to quantitatively assess the performance of the system.

To run the evaluation:

Bash

python evaluation/evaluate_rag.py