top of page

Data Curator

Acquire. Validate. Unify.

Description

Open-source library that downloads, validates, homogenizes and combines financial data from multiple providers into clean, point-in-time datasets ready for research. Every provider's fields are mapped to one consistent, readable naming scheme based on the US GAAP taxonomy, so you can switch data sources without changing your code. It can be configured entirely from an Excel workbook with no code required, run programmatically in Python as part of a larger system, or deployed with Docker.
 

Features

​​

  • Configurable from an Excel workbook, with no code required to download and curate data, or directly from a Python script. A Docker image is also available.

  • Provider-agnostic data model: readable, specific tag names based on the US GAAP taxonomy and homogenized across providers. Switch from one data provider to another without touching your code.

  • Automatic data validation: market and fundamental data are checked on download, and datasets that make no sense (such as a high price below the low) or can't guarantee point-in-time validity (such as amended statements) are discarded.

  • Custom calculated features: write your own feature functions in plain Python, with no need for NumPy or pandas (though you can use them if you prefer).

  • Flexible outputs: export to CSV or Parquet files, or return in-memory pandas DataFrames for further processing.

  • Fully extensible architecture: add your own data providers, feature combinations and output handlers on top of clear, stable interfaces.

  • AI-assistant friendly: readable, well-documented and tested code that works well with your favorite IDE and AI coding assistants.
     

Supported Data Providers
 

  • LSEG Workspace

  • Financial Modeling Prep

  • Yahoo! Finance

​

Everything you need to get started!
 

Data_ Curator.png
bottom of page