blobxiaoyao/quantalpha-nasdaq100-sample

🚀 QuantAlpha NASDAQ-100 Feature-Engineered Sample

This dataset provides a representative sample of the full QuantAlpha NASDAQ-100 dataset. It allows researchers and quantitative traders to explore the schema and feature richness of the NVDA dataset before accessing the full NASDAQ-100 universe.

---

📁 Dataset Contents

The sample consists of one Parquet file containing a randomly selected month of data from 2024 for NVIDIA (NVDA):

Ticker	Filename	Rows	Date Range
NVDA	NVDA_2024_month01.parquet	~21	January 2024

⚠️ Note: This is a limited sample. The full dataset includes all NASDAQ-100 constituents and multi-year historical coverage.

---

📊 Feature Overview

Each record contains 53 machine-learning-ready features, including:

• Trend Indicators: SMA ratios, MACD, ADX, Trend Persistence
• Momentum & Volatility: RSI, Stochastic Oscillator, ROC, Normalized ATR, Bollinger Band metrics
• Volume Metrics: On-Balance Volume (OBV), Volume Ratios
• Performance Metrics: Log Returns, 30-day Sharpe Ratio, 30-day Sortino Ratio
• Benchmark Analysis: Relative returns, Alpha, Beta vs. SPY and QQQ
• Market Microstructure: Price over Control (POC), Gap percentages, Z-scores

All features are cleaned, normalized, and free of look-ahead bias, making them ready for ML pipelines with XGBoost, LightGBM, or neural networks.

---

🛠 Usage

You can load the sample file directly into a Pandas DataFrame using fastparquet or pyarrow:

import pandas as pd

# Load the sample file
df = pd.read_parquet("NVDA_2024_month01.parquet")

# Inspect the data
print(df.info())
display(df.head())

📜 License

This sample is provided under the Creative Commons Attribution Non-Commercial 4.0 (CC BY-NC 4.0) license. For commercial licensing of the full NASDAQ-100 universe, please visit our Gumroad storefront.

---

📬 Contact & Support

If you have any questions about this dataset, licensing, or access to the full version, feel free to reach out:

📧 Email: quantalpha.global@gmail.com

Please note that this email is intended for dataset-related inquiries only.
We aim to respond within 1–2 business days.