realized-library 0.1.2

A Python package for high-frequency econometrics estimators

Realized Library

Realized Library is an open-source Python package for the computation of high-frequency econometrics estimators, such as variance, covariance and jump detection. It is designed as a modern, efficient, and production-ready continuation of the Oxford-Man Institute's original Realized Library.

The package focuses on performance and scalability, supporting large-scale high-frequency financial datasets (with trade, quote and time-sampled prices) up to the nanoseconds precision. The primary goal is to provide accurate, fast, and customizable estimators suitable for both academic research and quantitative analysis.

---

Background

This library is inspired by the Oxford-Man Institute of Quantitative Finance's original "Realized Library" (no longer maintained). It is designed for researchers, practitioners, and quantitative analysts who require fast and reliable estimations of variance, co-variance, jump and co-jumps in financial markets.

For reference to the underlying econometric methods, please consult the academic literature included in the original library or related publications in high-frequency financial econometrics. Since the website is no longer maintained, key information could be found on Wayback Machine trough this URL.

---

Key Features

• Efficient realized variance estimation from high-frequency price data
• Flexible resampling, subsampling and preaveraging utility with customizable parameters
• Designed for research, production pipelines, and large nanoseconds HFT datasets

---

Installation

pip install realized-library

Or, for development use:

git clone https://github.com/GabinTB/realized-library
cd realized-library
pip install -e .

The library requires Python 3.10.