I want to layer real-time sentiment intelligence over the cryptocurrency market data already flowing through my trading platform. The goal is simple: ingest live and historical crypto price feeds, fuse them with sentiment signals drawn from news, Twitter, Reddit or any reliable social-media/API source, and output a clean dataset plus a callable function that returns a sentiment score alongside each symbol’s price stream. The core market focus is 100 % crypto—no equities or forex—so feel free to rely on popular exchanges (Binance, Coinbase, Kraken, etc.) for price and volume while tapping natural-language APIs or custom NLP models for the social chatter. Python is preferred because the rest of my stack runs on Pandas, NumPy and a PostgreSQL back end, but I am open to alternative approaches if they integrate smoothly. What I expect at hand-off: • A well-documented script or micro-service that pulls market data and sentiment sources, performs the analysis, and writes results to my database or returns them as JSON. • Clear explanation of the sentiment methodology (lexicon, ML model, or hybrid) with performance metrics from a small back-test. • Setup instructions so I can deploy and schedule the job on my existing Ubuntu server. I will consider the task complete once I can schedule the sentiment job, query a coin (e.g., BTC-USDT) and receive a timestamped sentiment score that updates in line with incoming market data.