Tata Technologies Stock Move Explainer

What is this?

This is a tool that tries to answer a simple question: "Why did Tata Technologies' stock price move like that?"

Instead of guessing or predicting what the stock will do next, it looks backward at real evidence — the actual price history, recent news articles, and facts from the company's official annual reports — and connects the dots between them. It then explains, in plain language, what evidence lines up with a given price move.

This tool does not predict stock prices, and it does not give investment advice. It's an explanation tool, not a forecasting tool. Think of it like a research assistant that reads the news and the annual report for you, and shows its work.

How it works, roughly

1. Price data — Pulls historical stock price and volume data for Tata Technologies.
2. News — Pulls recent news headlines/articles about the company.
3. Annual reports — Reads PDF annual reports (you drop these in yourself) and pulls out relevant facts.
4. Knowledge graph — Connects all of this together into a graph, e.g. "price dropped on this date" → linked to → "news article about X" → linked to → "report mentions related risk factor Y."
5. Explanation — When you ask about a price move, it looks at what's connected to it in the graph and generates a plain-language explanation, citing that evidence — using Groq's free-tier AI model.

Everything shown to you should be traceable back to a real piece of evidence (a price data point, a news article, or a line from a report) — not made up by the AI.

Cost

This project is built to run at zero cost — it only uses free tools and free tiers of services (free stock data, free news feeds, and the free tier of Groq's AI models — no credit card required). No subscriptions or paid API keys required.

Getting started

1. Run the setup script to create a local Python environment and install everything needed:

   ./setup_venv.sh

2. Turn on the environment:

   source venv/bin/activate

3. Get a free Groq API key (no credit card required) from console.groq.com, then copy the example settings file and paste your key in:

   cp .env.example .env

4. If you have any annual report PDFs, drop them into data/raw/reports/.
5. Launch the app:

   streamlit run app/main.py

Project layout (what's in each folder)

• data/raw/reports/ — Where you manually place annual report PDFs.
• data/processed/ — Cleaned-up data the tool generates from the raw sources.
• src/ — The core logic (fetching prices, fetching news, reading PDFs, building the knowledge graph, generating explanations).
• app/ — The Streamlit web app you actually interact with.
• eval/ — Scripts for checking that explanations are accurate and well-grounded in evidence.