Fit spatial multilevel models and diagnose convergence
This is a package to estimate spatially-correlated variance components models/varying intercept models. In addition to a general toolkit to conduct Gibbs sampling in Python, the package also provides an interface to PyMC3 and CODA. For a complete overview, consult the walkthrough.
author: Levi John Wolf
institution: University of Bristol & University of Chicago Center for Spatial Data Science
preprint: on the Open Science Framework
This package works best in Python 3.5, but unittests pass in Python 2.7 as well. Only Python 3.5+ is officially supported.
To install, first install the Anaconda Python Distribution from Continuum Analytics. Installation of the package has been tested in Windows (10, 8, 7) Mac OSX (10.8+) and Linux using Anaconda 4.2.0, with Python version 3.5.
Once Anaconda is installed, spvcm can be installed using pip, the Python Package Manager.
pip install spvcm
To install this from source, one can also navigate to the source directory and use:
pip install ./
which will install the package from the target source directory.
To use the package, start up a Python interpreter and run: import spvcm.api as spvcm
Then, many differnet variance components model specificaions are available in:
spvcm.both spvcm.upper spvcm.lower
For more thorough directions, consult the Jupyter Notebook, using the sampler.ipynb, which is provided in the spvcm/examples directory.
Levi John Wolf. (2016). Gibbs Sampling for a class of spatially-correlated variance components models. University of Chicago Center for Spatial Data Science Technical Report.
Release history Release notifications
|Filename, size||File type||Python version||Upload date||Hashes|
|Filename, size spvcm-0.2.1.post1.tar.gz (6.2 MB)||File type Source||Python version None||Upload date||Hashes View hashes|