IGF python toolbox

Objectives

This package contains a set of Python utilities for :

• easy data loading ;
• descriptive statistics iteration ;
• implementation of econometric methods.

Organization

The repository is organized as follow :

• the s3 module contains a set of functions for importing or exporting files in various formats from or to an s3 bucket on the Nubonyxia platform;
• the stats_des module contains functions to complement those integrated by default in pandas for descriptive statistics (addition of totals, weighted statistics, verification of statistical confidentiality, etc.).
• the preprocessing module contains a set of classes that can be integrated into a sklearn.pipeline to perform various data transformation operations
• the model_selection module contains functions for training a prediction model or estimating a regression model
• the estimators module contains econometric models that can be integrated into a sklearn.pipeline
• the samples module contains methods for sample weighting and statistical inference
• the economics module contains functions to compute metrics, for industrial economics and concentration for instance
• the utils module contains a set of utility functions on which other functions in this module depend, or for calculating weighted statistics, for example

Installation

Package and dependencies

git clone <repo_url>
pip install -e igf_toolbox

The package can then be used like any other Python package.

Documentation

To visualize the documentation :

mkdocs build --port 5000

License

The package is licensed under the MIT License.