https://github.com/drsteve/pyforecasttools

Forecast Verification/Validation Tools in Python
https://github.com/drsteve/pyforecasttools

contingency-table forecast-verification model-validation python python-3 statistics

Last synced: 10 months ago
JSON representation

Forecast Verification/Validation Tools in Python

• Host: GitHub
• URL: https://github.com/drsteve/pyforecasttools
• Owner: drsteve
• License: bsd-3-clause
• Created: 2017-10-06T20:56:36.000Z (over 8 years ago)
• Default Branch: master
• Last Pushed: 2023-05-19T20:13:55.000Z (over 2 years ago)
• Last Synced: 2025-03-19T18:51:07.573Z (10 months ago)
• Topics: contingency-table, forecast-verification, model-validation, python, python-3, statistics
• Language: Python
• Homepage:
• Size: 385 KB
• Stars: 34
• Watchers: 5
• Forks: 11
• Open Issues: 4
• Metadata Files:
  • Readme: README.md
  • Changelog: CHANGELOG
  • License: LICENSE.md

Awesome Lists containing this project

README

          
# PyForecastTools

[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.1256921.svg)](https://doi.org/10.5281/zenodo.1256921)

[![Build Status](https://github.com/drsteve/PyForecastTools/workflows/CI/badge.svg?branch=master)](https://github.com/drsteve/PyForecastTools/actions?query=workflow%3ACI)

A Python module to provide model validation and forecast verification tools,

including a set of convenient plot functions. A selection of capabilites

provided by PyForecastTools includes:

 * Accuracy and bias metrics for continuous predictands

     - Unscaled/absolute measures

     - Relative measures

     - Scaled measures

 * 2x2 and NxN contingency table classes

     - Wide range of contingency table metrics and scores

     - Multiple methods of calculating confidence intervals on scores

 * Convenient plotting for visually comparing models and data

     - Quantile-Quantile plots

     - Taylor diagrams

     - ROC curves

     - Reliability diagrams

The module builds on the scientific Python stack (Python, Numpy, MatPlotLib)

and uses the dmarray class from SpacePy's datamodel.

SpacePy is available through the Python Package Index, MacPorts, and is under

version control at [github.com/spacepy/spacepy](https://github.com/spacepy/spacepy)

If SpacePy is not available a reduced functionality implementation of the class

is provided with this package.

PyForecastTools is available through the Python Package Index and can be installed

simply with

> pip install PyForecastTools --user

To install (local user), run

> python setup.py install --user

After installation, the module can then be imported (within a Python script or 

interpreter) by

> import verify

For help, please see the docstrings for each function and/or class.

Additional documentation is under development using Github pages at [drsteve.github.io/PyForecastTools](https://drsteve.github.io/PyForecastTools), and source for this is in the [docs folder](docs/).