Analysis of Trader Performance and Market Sentiment

This repository contains a data analysis pipeline investigating the relationship between trader performance (derived from trading history data on Hyperliquid) and broader cryptocurrency market sentiment (measured by the Bitcoin Fear and Greed Index).

Project Overview

The core objective of this project is to explore whether market sentiment (fear vs. greed) correlates with or influences:

1. Trader Profitability (PnL): How average daily profits and losses vary under different sentiment classifications.
2. Trading Activity & Volume: Whether extreme fear or greed leads to higher daily trading volumes and trade counts.
3. Transaction Costs: How total daily transaction fees behave under various sentiment regimes.

---

Directory Structure

• Analysis.ipynb: Main Jupyter notebook containing the data preprocessing, exploratory data analysis, correlation studies, and visualization.
• historical_data.csv: A comprehensive log of historical trades, including price, size, side (BUY/SELL), timestamp, closed PnL, and fees.
• fear_greed_index.csv: Historical daily records of the Bitcoin Fear and Greed Index, classifying days into categories like Extreme Fear, Fear, Greed, and Extreme Greed.
• requirements.txt: Python package requirements.
• read_notebook.py: A utility script to inspect Jupyter notebook cells.

---

Dataset Details

1. Trader Historical Data (historical_data.csv)

This dataset contains trade-level execution details for a trader's account. Key columns include:

Column	Description	Data Type
Account	The trader's wallet address	String (Hexadecimal)
Coin	Token code or identifier	String
Execution Price	Execution price of the trade	Float
Size Tokens	Quantity of tokens traded	Float
Size USD	Transaction size in USD	Float
Side	Trade direction (BUY or SELL)	String
Timestamp IST	Trade execution time (IST, format: %d-%m-%Y %H:%M)	String / Datetime
Closed PnL	Realized profit and loss in USD (populated for closing trades)	Float
Fee	Transaction fees paid	Float
Order ID	Unique identifier for the order	Integer

2. Bitcoin Fear and Greed Index (fear_greed_index.csv)

This dataset represents the daily market sentiment for Bitcoin. Key columns include:

Column	Description	Data Type
timestamp	Unix timestamp of the index record	Integer
value	Sentiment score ranging from 0 (Extreme Fear) to 100 (Extreme Greed)	Integer
classification	Qualitative sentiment category (e.g., Fear, Extreme Fear, Greed, Extreme Greed)	String
date	Calendar date (format: YYYY-MM-DD)	String / Datetime

---

Installation and Setup

Prerequisites

Make sure you have Python 3.8+ installed on your system.

Installing Dependencies

Install the required libraries listed in requirements.txt:

pip install -r requirements.txt

Running the Analysis

To view and execute the analysis pipeline:

1. Start the Jupyter notebook server in the project directory:

   jupyter notebook

2. Open Analysis.ipynb and run all cells.

---

Analysis Pipeline

1. Data Loading and Preprocessing

• Type Conversion: Converts time columns to datetime objects. Converts string numeric values to proper floats/ints, coercing errors to NaN.
• Daily Aggregation: Groups the detailed trade data from historical_data.csv by date to calculate:
  • total_pnl: Sum of realized profits/losses per day.
  • total_volume: Sum of size in USD per day.
  • total_trades: Count of trades executed per day.
  • total_fees: Sum of transaction fees paid per day.
• Merging: Performs an inner join of the daily trading statistics and the daily Fear & Greed index using the date.

2. Exploratory Data Analysis (EDA)

• Daily PnL by Sentiment: Generates a bar plot showing the average daily PnL across different sentiment classifications (e.g., Fear, Extreme Fear, Greed, Extreme Greed).
• Daily Volume by Sentiment: Generates a bar plot illustrating average trading volume across different market sentiments.

3. Correlation Analysis

• Computes a correlation matrix between total_pnl, total_volume, total_trades, total_fees, and the Fear & Greed index value.
• Renders a Seaborn heatmap representing these correlations to identify linear relationships between sentiment score and trading metrics.

4. Key Findings & Insights

• Trading Volume & Sentiment: Extreme market sentiment (extreme greed/fear) often corresponds to elevated volumes.
• Profitability (PnL): Highlights whether trading profits increase during bullish sentiment (Greed) or defensive phases (Fear).
• Correlation Strength: Indicates the level of linear dependency between sentiment metrics and fees or transaction counts.