Building a Crypto Fear and Greed Index Tracker with Python

#python #tutorial #webdev #programming

The Crypto Fear & Greed Index measures sentiment from 0 (extreme fear) to 100 (extreme greed). Build a tracker combining this with price data for better trading signals.

Alternative.me API

import requests, statistics
from datetime import datetime

class FearGreedTracker:
    def __init__(self):
        self.fg = "https://api.alternative.me/fng/"
        self.cg = "https://api.coingecko.com/api/v3"

    def current(self):
        d = requests.get(f"{self.fg}?limit=1").json()["data"][0]
        return {"value":int(d["value"]),"label":d["value_classification"]}

    def history(self, days=90):
        return [{"value":int(d["value"]),"label":d["value_classification"],
                 "date":datetime.fromtimestamp(int(d["timestamp"])).strftime("%Y-%m-%d")}
                for d in requests.get(f"{self.fg}?limit={days}").json()["data"]]

    def btc_prices(self, days=90):
        r = requests.get(f"{self.cg}/coins/bitcoin/market_chart",
            params={"vs_currency":"usd","days":days})
        return {datetime.fromtimestamp(p[0]/1000).strftime("%Y-%m-%d"):p[1]
                for p in r.json()["prices"]}

Correlation Analysis

    def correlate(self, days=90):
        fg, prices = self.history(days), self.btc_prices(days)
        paired = [{"fg":e["value"],"price":prices[e["date"]]}
                  for e in fg if e["date"] in prices]
        sigs = {"fear":[],"greed":[]}
        for i, p in enumerate(paired):
            if i+7 < len(paired):
                ret = (paired[i+7]["price"]-p["price"])/p["price"]*100
                if p["fg"] <= 20: sigs["fear"].append(ret)
                elif p["fg"] >= 80: sigs["greed"].append(ret)
        for k, v in sigs.items():
            if v: print(f"{k}: avg 7d={statistics.mean(v):+.2f}%, win={sum(1 for r in v if r>0)/len(v)*100:.0f}%")

    def social_sentiment(self):
        r = requests.get("https://www.reddit.com/r/cryptocurrency/hot.json",
            headers={"User-Agent":"Bot/1.0"})
        posts = r.json()["data"]["children"]
        bull = sum(1 for p in posts if any(w in p["data"]["title"].lower() for w in ["bull","moon","pump"]))
        bear = sum(1 for p in posts if any(w in p["data"]["title"].lower() for w in ["bear","crash","dump"]))
        t = bull+bear or 1
        return {"bull":bull/t*100,"bear":bear/t*100}

    def composite(self):
        fg, soc = self.current(), self.social_sentiment()
        score = fg["value"]*0.6 + soc["bull"]*0.4
        print(f"Composite: {score:.0f}/100 | F&G:{fg['value']} | Social:{soc['bull']:.0f}% bull")

t = FearGreedTracker()
print(t.current())
t.correlate(90)
t.composite()