dlmhelper 1.0.1

This package includes functions and custom data types to help with performing dynamic linear model fits and analyzing the results.

dlmhelper

This package includes functions and custom data types to help with performing and analyzing dynamic linear model fits.

The initial version of this code was developed for the analysis performed in 'Zonal variability of methane trends derived from satellite data' (Hachmeister et al., 2023 ; DOI: 10.5194/acp-24-577-2024).

Version v0.1.0 of this package is also used by the supplementary example code provided for the above mentioned paper, available on Zenodo.

Requirements

• numpy
• matplotlib
• statsmodels
• tabulate

Version

v1.0.1

Installation

Install from PyPI using:

pip install dlmhelper

Usage

Look at the Nile or Manua Loa example from the examples folder for a basic introduction.

Documentation is available here

Changelog

v1.0.1

• Fixed occurence of negative variances by adding a "burn-in" period, where variances are discarded (i.e., set to NaN).
• Added two new helper functions for DLMResult objects that allow for easy access to the fitted data and timestamps (get_fit, get_time)

Known issues

• Sometimes variances (e.g., trend_cov) are negative. This is a numerical artifact from the underlying fitting routine used by the statsmodels package