This project investigates how Bitcoin market sentiment (the Fear & Greed Index) influences trader profitability and risk behaviour. It merges day-by-day sentiment with real trade-level records from the Hyperliquid exchange, then analyses profitability, risk and trading activity across Fear, Neutral and Greed regimes.
The goal: find out whether market psychology is a usable signal for better entries, position sizing and risk management in crypto trading.
1. Bitcoin Fear & Greed Index (fear_greed_index.csv) — included in this repo
- 2,644 daily readings, Feb 2018 → May 2025
- Columns:
timestamp,value(0–100),classification,date - Sentiment split: ~49% Fear days · ~36% Greed days, average index ≈ 47/100
2. Hyperliquid Trader Data (historical_data.csv) — trade-level
- Account, coin, execution price, size (tokens / USD), side, timestamp, closed PnL, fees
- Reflects real trader behaviour in live crypto markets
Sentiment over seven years — the market swings hard between fear and greed:
How often the market sat in each mood:
- Clean both datasets — consistent dates, numeric coercion, de-duplication.
- Integrate — merge the Fear & Greed Index onto trader records by date.
- EDA by regime — compare average & total closed PnL, win-rate, trade size and activity across Fear / Neutral / Greed.
- Statistics — correlation and group tests to check whether differences are meaningful.
- Trader performance varies materially across sentiment regimes — the mood of the market coincides with measurable shifts in profitability.
- Risk behaviour differs in Fear vs Greed — position sizing and activity are not constant across regimes.
- Sentiment is a usable context signal — pairing the Fear & Greed Index with execution discipline can inform smarter risk management.
(Full figures, tables and statistical tests are in the notebook.)
git clone https://github.com/Namanjain723/Bitcoin-Sentiment-Trader-Analysis.git
cd Bitcoin-Sentiment-Trader-Analysis
pip install -r requirements.txt
jupyter notebook trader_sentiment_analysis.ipynb
fear_greed_index.csvships with the repo. To reproduce the merge/PnL analysis end-to-end, place the Hyperliquidhistorical_data.csvin the project folder (it is large and sourced from the exchange, so it is not committed here).
Python · Pandas · NumPy · SciPy · Matplotlib · Jupyter Notebook
Naman Jain — Data Analyst & AI Developer 🌐 Portfolio · ✉️ namancric01@gmail.com · 🐙 @Namanjain723