powerlaw

2.0.0 verified Fri May 01 auth: no python

A toolbox for testing if a probability distribution fits a power law, including fitting power laws, comparing to other distributions via likelihood ratio tests, and plotting. Current version is 2.0.0 (released 2023-12-31), with major refactor from v1.x. Release cadence is irregular.

pip install powerlaw

Common errors

error AttributeError: module 'powerlaw' has no attribute 'Fit' ↓

cause Old code that imported a submodule incorrectly or used an outdated import pattern.

fix

Use import powerlaw and then fit = powerlaw.Fit(data).

error TypeError: Fit() got an unexpected keyword argument 'xmin' ↓

cause The API changed in v2.0; `xmin` is no longer a direct argument to `Fit`; use `Fit(data, xmin=value)` may still work but check docs.

fix

Use fit = powerlaw.Fit(data, xmin=...) and refer to v2.0 docs; if using v2.0, the argument is still supported but check for other changes.

error ValueError: Data must be positive ↓

cause Power law fitting requires all data points > 0.

fix

Filter your data: data = data[data > 0] before fitting.

Warnings

breaking Version 2.0.0 is a major refactor with API changes. The previous API (e.g., `powerlaw.Fit` returning a `Result` object with `.power_law.alpha`) has been simplified. Ensure you update code and refer to new documentation. ↓

fix Review the new API: `Fit` now directly returns the fitted parameters (alpha, xmin) as attributes. Use `fit.distribution_compare(...)` for comparisons.

deprecated Old plotting functions from v1.x (e.g., `fit.plot_pdf()`) may be removed. Use `fit.plot_ccdf()` or `fit.plot_pdf()` as per new API. ↓

fix See the documentation for new plotting methods; they have similar names but may have different signatures.

gotcha Data with values <= 0 will cause errors. Power law distributions are defined for positive values only. ↓

fix Ensure your data is strictly positive (e.g., remove zeros and negative values before fitting).

Imports

powerlaw
wrong
```
from powerlaw import powerlaw
```
correct
```
import powerlaw
```
Library is imported directly as 'powerlaw' module; no submodule

Quickstart

Fit a power law to data and compare to alternative distributions.

import powerlaw
import numpy as np
# Generate sample data
x = np.random.pareto(2.5, 1000)
# Fit power law
fit = powerlaw.Fit(x)
# Print alpha and xmin
print(fit.alpha, fit.xmin)
# Compare to exponential
R, p = fit.distribution_compare('power_law', 'exponential')