PrimeTrade.ai: Market Sentiment vs. Trader Performance Analysis

This repository contains the data science assignment analyzing the relationship between Bitcoin market sentiment (Fear & Greed Index) and historical trader performance from the Hyperliquid perpetual DEX.

The analysis is based on a 2-year overlapping dataset (May 1, 2023 – May 1, 2025) comprising 211,224 trade executions across 246 assets (with asset-specific focus on BTC).

---

🚀 How to Run the Project

1. Clone the Repository

git clone https://github.com/Yashm2610/PrimeTrade.ai.git
cd PrimeTrade.ai

2. Install Dependencies

Make sure you have Python installed, then run:

pip install pandas numpy matplotlib seaborn scipy streamlit

3. Download the Datasets

Run the downloader script to fetch the Hyperliquid trader history and Bitcoin Fear & Greed dataset from Google Drive:

python download_data.py

4. Run the Statistical Analysis

Run the core analysis script. This script cleans and aligns the timestamps (using Timestamp IST to fix the CSV scientific notation precision loss), computes Pearson correlations, runs One-way ANOVA tests, saves statistics to analysis_output.txt, and generates static plots in the charts/ directory:

python analysis.py

5. Launch the Interactive Web Dashboard

Run the Streamlit application to open the interactive, glassmorphic analytics dashboard in your browser:

streamlit run app.py

The dashboard allows filtering by target asset (Overall vs. BTC-only), adjusting the Fear & Greed range on the fly, rendering real-time KPI metrics, showing dynamic ANOVA/correlation tests, and exploring/downloading the data.

---

📊 Key Trading Insights

1. Profitability is Counter-Cyclical: Daily trader profits (Closed PnL) are negatively correlated with sentiment. Traders make the most during Extreme Fear ($52.8k/day average) and the least during Greed ($11.1k/day average).
2. Volatility Spikes Trading Activity: Daily trade count and volume are strongly negatively correlated with sentiment ($p < 0.001$). Trading counts during Extreme Fear are 4.4x higher than during Greed.
3. Traders are Contrarians: Hyperliquid traders actively buy the dip in Fear (Long Open Ratio rises to 66.7% overall, and 78.2% for BTC-only) and short/hedge in Greed (Long Open Ratio falls to 49.2% in Extreme Greed).
4. Win Rate vs. PnL Asymmetry: In Extreme Fear, traders show their lowest win rate (65.4%) but highest daily PnL ($52.8k). This implies that they take multiple small losses attempting to catch bottoms, but their winning trades yield massive, outsized returns.

---

📂 Repository Structure

• download_data.py: Programmatic downloader script for Google Drive hosted datasets.

• analysis.py: Main processing script that runs the statistical tests and generates static charts.

• app.py: Streamlit-based interactive web dashboard with custom dark-theme glassmorphism.

• report.md: Deep-dive executive analysis report detailing findings, tables, and recommendations.

• analysis_output.txt: Captured numeric logs and statistical test outputs of the run.

• charts/: Folder containing saved visualization figures.

• bitcoin_market_sentiment.csv: The processed Bitcoin Fear & Greed index historical data.

• Added dynamic charts and statistical testing features