Data Analyst · advanced
Real-Time Market Sentiment & Price Correlation Tracker
A live pipeline that fetches crypto/stock prices AND news sentiment, visualizing the correlation in real-time.
Project overview
Proves you can handle live, streaming data and unstructured text — skills that distinguish top-tier analysts from those who only work with static CSV files.
Real-world use case
Fintech startups and hedge funds that need to monitor market volatility and public perception in real-time to make split-second decisions.
Tools and technologies
- Python
- WebSocket
- NLTK/TextBlob
- Streamlit
- AWS Lambda
Skills you can demonstrate
API Integration, Real-time Processing, Sentiment Analysis, Data Viz
Build roadmap
- Set up your data sources
Get API keys for CoinGecko (free) and NewsAPI (free tier) — or use Reddit's PRAW for social sentiment. - Build the price streamer
Poll CoinGecko's REST API every 30s (free tier is REST-only). For true WebSocket streaming, use Binance or Coinbase. Store tick data in a rolling window (pandas DataFrame or Redis). - Build the news fetcher
Every 5 minutes, pull the latest headlines mentioning your asset. Deduplicate by URL. - Score sentiment
Start with VADER for speed, upgrade to a fine-tuned FinBERT for accuracy. Average sentiment over a rolling window. - Build the correlation view
Plot price and sentiment on the same Streamlit chart with dual y-axes. Add a Pearson correlation coefficient that updates live. - Deploy to production
Host the backend on AWS Lambda (triggered by EventBridge) and the dashboard on Streamlit Cloud. Include a public URL in your portfolio.
Ways to extend the project
Build a "Predictive Alert" system using an LSTM model or AI agent that flags unusual sentiment spikes before they impact the price.
How to showcase your work
- GitHub: Architecture diagram showing the real-time data flow.
- LinkedIn: Live link to the hosted dashboard.
- Resume: Mention real-time API processing and NLP.