IITKGP/Nifty500_companies_5_years_data

Dataset Card for Dataset_of nifty500

Dataset Details

Dataset Description

This dataset contains 5 years of daily historical stock market data for companies listed in the Nifty 500 index on the National Stock Exchange of India (NSE). It provides essential end-of-day price metrics, including Open, High, Low, and Close (OHLC) prices, as well as trading Volume and a Notes column for potential annotations. The dataset spans approximately from mid-2021 to mid-2026, serving as a robust foundational dataset for financial machine learning (ML), artificial intelligence (AI) forecasting models, and quantitative analysis.

• Curated by: Open-source community / Shreyash, IIT Kharagpur
• Funded by [optional]: N/A
• Shared by [optional]: Open-source contributors
• Language(s) (NLP): English (en)
• License: MIT

Dataset Sources [optional]

• Repository: https://huggingface.co/datasets/ShreyashAutomation/Dataset_of_nifty500
• Paper [optional]: N/A
• Demo [optional]: N/A

Uses

Direct Use

This dataset is intended for researchers, quants, and data scientists looking to build and test:

• Time-series forecasting models: Predicting future stock prices or market movements.
• Algorithmic Trading Strategies: Backtesting trading logic on historical Nifty 500 data.
• Financial AI projects: Feature engineering for larger ML pipelines or generative AI applications tailored to the Indian stock market.
• Market Analysis: Analyzing volume trends, volatility, and sector-wide momentum.

Out-of-Scope Use

• Direct Financial Advice: Models trained on this dataset should not be used as standalone financial advisors for live trading without professional oversight.
• High-Frequency Trading: The data provides daily intervals, making it unsuitable for intraday, tick-by-tick, or high-frequency trading (HFT) strategy development.

Dataset Structure

The dataset is provided in a single CSV format (nifty500_5yr_data.csv). The columns are structured as follows:

• Date: The trading date (YYYY-MM-DD).
• Ticker: The stock symbol for the Nifty 500 company (e.g., 360ONE, ECLERX).
• Open: The opening price of the stock on the given date.
• High: The highest price reached during the trading session.
• Low: The lowest price reached during the trading session.
• Close: The closing price at the end of the trading session.
• Volume: The total number of shares traded during the day.
• Notes: Additional annotations or metadata related to the trading day (if applicable).

Dataset Creation

Curation Rationale

The Nifty 500 index represents the top 500 companies based on full market capitalization from the eligible universe of the Indian equity market. This dataset was aggregated to provide a clean, standardized, and easily accessible historical record to facilitate AI and ML research specifically focused on the Indian financial landscape.

Source Data

Data Collection and Processing

Data was aggregated from public market data providers spanning a 5-year period. Processing steps included organizing the data by Date and Ticker, structuring OHLC values to standardized floating-point numbers, and ensuring a uniform CSV layout for seamless integration with libraries like pandas and datasets.

Who are the source data producers?

The original underlying price and volume data is generated by the National Stock Exchange of India (NSE) and collected via standard financial APIs or market aggregators.

Annotations [optional]

Annotation process

The Notes column allows for custom annotations (e.g., marking stock splits, dividend dates, or major news events). Currently, this is heavily sparse or auto-generated based on the collection script.

Personal and Sensitive Information

This dataset contains no personal, private, or sensitive information. All data represents publicly traded corporate entities and macro-market movements.

Bias, Risks, and Limitations

• Survivorship Bias: Depending on how the Nifty 500 components were tracked over the 5 years, the dataset might contain survivorship bias if delisted companies were retroactively removed from the historical data.
• Corporate Actions: Users should verify if the OHLC prices are adjusted for corporate actions like stock splits and dividends, as raw unadjusted data can cause massive artificial price drops/spikes in ML models.
• Market Regime Changes: 5 years of data captures specific economic conditions (e.g., post-COVID recovery). AI models trained strictly on this period may struggle to generalize to drastically different macroeconomic regimes.

Recommendations

Users should ensure they normalize the data before passing it into neural networks and consider incorporating external macroeconomic indicators (like inflation rates, interest rates, or Nifty 50 index data) to provide broader context to their AI models.

Citation [optional]

BibTeX:

@misc{nifty500_5yr_dataset,
  author = {Shreyash Mishra},
  title = {Nifty 500 5-Year Historical OHLCV Dataset},
  year = {2026},
  publisher = {Hugging Face},
  howpublished = {\url{[https://huggingface.co/datasets/ShreyashAutomation/Dataset_of_nifty500](https://huggingface.co/datasets/ShreyashAutomation/Dataset_of_nifty500)}}
}