Installing via pip

To install the PySD package from the Python package index into an established Python environment, use the pip command:

pip install pysd

Installing from source

To install from the source, clone the project with git:

git clone

Or download the latest version from the project webpage:

In the source directory use the command

python install

Required Dependencies

PySD is built on python 2.7, and may not work as advertized on 3.x.

PySD calls on the core Python data analytics stack, and a third party parsing library:

  • Numpy
  • Scipy
  • Pandas
  • Matplotlib
  • Parsimonious

These modules should build automatically if you are installing via pip. If you are building from the source code, or if pip fails to load them, they can be loaded with the same pip syntax as above.

Optional Dependencies

These Python libraries bring additional data analytics capabilities to the analysis of SD models:

  • PyMC: a library for performing Markov chain Monte Carlo analysis
  • Scikit-learn: a library for performing machine learning in Python
  • NetworkX: a library for constructing networks
  • GeoPandas: a library for manipulating geographic data

Additionally, the System Dynamics Translator utility developed by Robert Ward is useful for translating models from other system dynamics formats into the XMILE standard, to be read by PySD.

These modules can be installed using pip with syntax similar to the above.

Additional Resources

The PySD Cookbook contains a recipe on Installation_and_Setup that can help you get set up with both python and PySD.