CryptoSense 🚀

AI-powered meme coin trend predictor that turns social media noise into a single actionable signal.

What it does

CryptoSense collects real-time data from Reddit, NewsAPI, StockTwits, and CoinGecko, runs sentiment analysis, and produces a Hype Score (0–100) per coin — telling you whether a meme coin is Rising, Flat, or Falling.

How it works

Reddit + NewsAPI + StockTwits + CoinGecko
              ↓
     VADER Sentiment Analysis
     + Custom Crypto Lexicon
              ↓
     6-Signal Hype Score Formula
              ↓
     Live Dashboard (updates every 60s)

Hype Score Formula

Hype Score = 0.25 × Mentions
           + 0.20 × Sentiment
           + 0.20 × Volume
           + 0.15 × Price Momentum
           + 0.10 × Engagement
           + 0.10 × News Count

---

Tech Stack

Layer	Tools
Language	Python 3.10+
Sentiment	VADER (nltk) + Custom Crypto Lexicon
Data	Pandas, NumPy
Backend API	Flask
Frontend	HTML, CSS, JavaScript
Data Sources	Reddit, NewsAPI, StockTwits, CoinGecko

---

Getting Started

1. Clone the repo

git clone https://github.com/yourusername/cryptosense.git
cd cryptosense

2. Install dependencies

pip install requests nltk pandas numpy flask flask-cors
python -c "import nltk; nltk.download('vader_lexicon')"

3. Add your API keys

Open scorer.py and paste your keys in Section 1 at the top:

NEWS_API_KEY        = "your_newsapi_key"
CRYPTOPANIC_API_KEY = "your_cryptopanic_key"  # optional

Free API keys:

• NewsAPI → newsapi.org/register
• CryptoPanic → cryptopanic.com/api
• CoinGecko → no key needed
• StockTwits → no key needed

4. Run

Open 3 terminals:

# Terminal 1 — scorer (keeps running, updates every 60s)
python scorer.py

# Terminal 2 — API server
python api.py

# Terminal 3 — frontend
cd frontend
python -m http.server 5500

Open browser at http://localhost:5500

---

Project Structure

cryptosense/
├── scorer.py                  # AI scoring engine — runs sentiment + hype formula
├── api.py                     # Flask API server — serves scored_coins.csv
├── requirements.txt           # Python dependencies
├── scored_coins.csv           # auto-generated by scorer.py
├── coins_cache.csv            # auto-generated CoinGecko cache
└── frontend/
    ├── index.html             # main entry point
    ├── css/
    │   ├── base.css
    │   ├── components.css
    │   └── layout.css
    ├── js/
    │   ├── api.js             # fetches data from Flask API
    │   ├── charts.js          # chart rendering
    │   ├── data.js            # demo data + coin metadata
    │   ├── router.js          # single page navigation
    │   └── utils.js           # helper functions
    └── pages/
        ├── alerts.html        # hype alerts page
        ├── coin detail.html   # individual coin drill-down
        ├── overview.html      # market overview
        ├── reddit-feed.html   # news + social feed
        ├── sentiment.html     # sentiment breakdown
        └── trending.html      # trending coins

Coins Tracked

Top 100 coins by market cap — fetched dynamically from CoinGecko. Includes DOGE, SHIB, PEPE, WIF, BONK and 95 more.

---

Built at

Dev Hub 1.0 Hackathon — Track 1: Crypto & Trend Intelligence

License

MIT