properscoring
Proper scoring rules in Python
Decision gist · record as of 2026-08-14
Yes, if you need strictly proper scoring rules for probabilistic forecast evaluation and can tolerate an abandoned package. The core algorithms are scientifically sound and the library has no known vulnerabilities. However, verify compatibility with your Python version before relying on it in production; consider seeking an actively maintained alternative if you need ongoing support or updates.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires NumPy (1.8 or later) and SciPy; Numba is optional but highly encouraged for performance on crps_ensemble and threshold_brier_score.
- Installation is straightforward with no runtime dependencies, though the package is abandoned—last release was 2015-11-12 and last commit 2023-03-09.
- It targets Python 2 and early Python 3 versions; compatibility with modern Python is unverified.
License · maintenance · safety
Apache (permissive) — Licensed under Apache License 2.0 (permissive), allowing commercial and private use with minimal restrictions, though you must include a copy of the license and state any modifications.
last release 2015-11-12 (3928 days) · last repo commit 2023-03-09 · 188 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 167,190 downloads/mo, #10,478 on PyPI
Alternatives
Verify before relying
pip install properscoring
import properscoring as ps
# Calculate CRPS for a Gaussian forecast
ps.crps_gaussian(0, mu=0, sig=1)- Whether the package works reliably on Python 3.6 and later versions despite classifiers only listing up to 3.5.
- Current performance characteristics and whether Numba speedups (e.g., 20x) still apply with modern versions.
- Whether the package's core algorithms remain scientifically current for modern forecast evaluation workflows.
What it is and what it does
Properscoring is a Python library for computing strictly proper scoring rules—evaluation metrics that cannot be artificially improved through hedging—making them fair and reliable for assessing the accuracy of probabilistic forecasts. It implements Continuous Ranked Probability Score (CRPS) and Brier Score in multiple forms: CRPS can be calculated exactly for Gaussian distributions, via numerical integration for arbitrary distributions, from finite ensemble samples, or weighted by probability density values; Brier Score is provided for binary forecasts and threshold exceedances. All functions handle multi-dimensional arrays and missing values (NaN).
The library was created by researchers at The Climate Corporation and is particularly useful in weather forecasting and other domains where probabilistic predictions need rigorous evaluation. It requires external dependencies and optional Numba support for performance improvements on ensemble-based calculations. However, the package is no longer actively maintained—the last release and commit occurred years ago—so compatibility with recent Python versions may be limited.
Use it for
- Evaluate weather forecast model accuracy by computing CRPS scores across many observations and comparing to baseline forecasts.
- Calculate skill scores to measure forecast improvement relative to a baseline, normalized to a 0–1 scale.
- Score ensemble forecasts from climate or weather models using finite samples without requiring the underlying distribution.
- Compute Brier scores for binary probabilistic predictions or threshold exceedances.
- Batch-process large arrays of forecast–observation pairs to generate aggregate performance metrics.
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, if you need strictly proper scoring rules for probabilistic forecast evaluation and can tolerate an abandoned package.
The core algorithms are scientifically sound and the library has no known vulnerabilities. However, verify compatibility with your Python version before relying on it in production; consider seeking an actively maintained alternative if you need ongoing support or updates.
Install
properscoring on PyPI
Before you install
Installation is straightforward with no runtime dependencies, though the package is abandoned—last release was 2015-11-12 and last commit 2023-03-09. It targets Python 2 and early Python 3 versions; compatibility with modern Python is unverified.
Requires NumPy (1.8 or later) and SciPy; Numba is optional but highly encouraged for performance on crps_ensemble and threshold_brier_score.
License in practice
Licensed under Apache License 2.0 (permissive), allowing commercial and private use with minimal restrictions, though you must include a copy of the license and state any modifications.
Quickstart
pip install properscoring
import properscoring as ps
# Calculate CRPS for a Gaussian forecast
ps.crps_gaussian(0, mu=0, sig=1)Verify before relying
- Whether the package works reliably on Python 3.6 and later versions despite classifiers only listing up to 3.5.
- Current performance characteristics and whether Numba speedups (e.g., 20x) still apply with modern versions.
- Whether the package's core algorithms remain scientifically current for modern forecast evaluation workflows.
Package facts
| License | Apache permissive |
| Python support | Not specified |
| Install friction | Low. Pure-Python wheel |
| Runtime dependencies | None |
| Maintenance | Abandoned 3,928 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 167,190 / month, #10,478 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | Development Status :: 3 - AlphaIntended Audience :: Science/ResearchLicense :: OSI Approved :: Apache Software LicenseOperating System :: OS IndependentProgramming Language :: PythonProgramming Language :: Python :: 2Programming Language :: Python :: 2.7Programming Language :: Python :: 3Programming Language :: Python :: 3.4Programming Language :: Python :: 3.5Topic :: Scientific/Engineering |
Evidence: properscoring-0.1-py2.py3-none-any.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “probabilistic forecast evaluation”
- properscoringProperscoring provides strictly proper scoring rules for evaluating…
- hierarchicalforecastReconciles hierarchical time-series forecasts across multiple levels…
- gluontsGluonTS provides deep learning models for probabilistic time series…
Give your agent the search over MCP, or paste the wish link into any chat.
More Scientific/Engineering packages
NumPy provides an N-dimensional array object and a comprehensive suite of mathematical, linear algebra, Fourier transform, and random number functions for scientific computing in Python.
pandas provides fast, flexible data structures (Series and DataFrame) for loading, cleaning, transforming, and analyzing labeled or relational data in Python.
scipy provides numerical algorithms for mathematics, science, and engineering—including optimization, integration, linear algebra, Fourier transforms, signal and image processing, and ODE solvers—built on numpy arrays.
scikit-learn provides a comprehensive Python library for supervised and unsupervised machine learning, including classification, regression, clustering, dimensionality reduction, and model evaluation tools built on NumPy and SciPy.
Install it if you need to train, evaluate, or deploy supervised or unsupervised learning models.
dill extends Python's pickle module to serialize and deserialize a much wider range of Python objects, including functions, lambdas, classes, and interpreter sessions, to byte streams for storage or network transmission.
Multiprocess is an enhanced fork of Python's standard multiprocessing library that uses dill for better serialization, allowing you to spawn processes with a threading-like API and share complex objects between them.
Install it if you use multiprocessing and encounter pickle serialization limits with lambdas or complex objects.
See also hierarchicalforecast · ngboost · skope-rules · AI-WQ-package · tfp-nightly · tensorflow-probability · problog · pyjpt · zxcvbn · cvss