Neural Network Powered Sentiment Analysis
For research purposes only. Not financial advice.
Fear & Greed Index data provided by Alternative.me Crypto Fear & Greed Index
How it works:
This tool runs a small neural network directly in your browser. It is trained on historical Bitcoin market data and Fear & Greed Index values to classify market sentiment into five categories: Extreme Fear, Fear, Neutral, Greed, and Extreme Greed.
Training data:
The model is trained on a static historical dataset (2018–2023) compiled from public sources including Alternative.me (Fear & Greed Index) and Yahoo Finance (price/volume), published on Kaggle by Adil Bhatti.
Live data (optional):
You can fetch current Bitcoin price and 24h volume from CoinGecko to run a prediction for “right now”. Live data is used only for prediction (inference), not for training.
Model & tech:
Built with p5.js (UI), ml5.js / TensorFlow.js (neural network), and PapaParse (CSV parsing). Everything runs client-side — no server, no accounts, no tracking.
Disclaimer:
For research purposes only. Not financial advice. Crypto markets are volatile and unpredictable — do not use this tool for investment decisions.
Author:
Rene Mathis — AI Product & Experience Design · Portfolio · LinkedIn