Population Activity Modeller

PAM is a python library for population activity sequence modelling. Example use cases:

• Read an existing population then write to a new format.
• Modify an existing population, for example to model activity locations.
• Create your own activity-based model.

PAM supports common travel and activity formats, including MATSim.

Activity Sequences?

Population activity sequences (sometimes called activity plans) are used to model the activities (where and when people are at home, work, education and so on) and associated travel of a population:

Activity sequences are used by transport planners to model travel demand, but can also be used in other domains, such as for virus transmission or energy use modelling.

Brief History

PAM was originally built and shared to rapidly modify existing activity models to respond to pandemic lock-down scenarios.

This functionality used a read-modify-write pattern. Where modifications are made by applying policies. Example policies might be (a) infected persons quarantine at home, (b) only critical workers travel to work, and (c) everyone shops locally.

Features

Activity Modelling

In addition to the original read-modify-write pattern and functionality, PAM has modules for:

• location modelling
• discretionary activity modelling
• mode choice modelling
• facility sampling
• vehicle ownership

More generally the core PAM data structure and modules can be used as a library to support your own use cases, including building your own activity-based model.

MATSim

PAM fully supports the MATSim population/plans format. This includes vehicles, unselected plans, leg routes and leg attributes. A core use case of PAM is to read-modify-write experienced plans from MATSim. This can allow new MATSim scenarios to be "warm started" from existing scenarios, significantly reducing MATSim compute time.

Documentation

For more detailed instructions, see our documentation.

Installation

To install PAM, we recommend using the mamba package manager:

As a user

git clone git@github.com:arup-group/pam.git
cd pam
mamba create -n pam -c conda-forge -c city-modelling-lab --file requirements/base.txt
mamba activate pam
pip install --no-deps .

As a developer

git clone git@github.com:arup-group/pam.git
cd pam
mamba create -n pam -c conda-forge -c city-modelling-lab --file requirements/base.txt --file requirements/dev.txt
mamba activate pam
pip install --no-deps -e .

For more detailed instructions, see our documentation.

Contributing

There are many ways to make both technical and non-technical contributions to PAM. Before making contributions to the PAM source code, see our contribution guidelines and follow the development install instructions.

If you are using pip to install PAM instead of the recommended mamba, you can install the optional test and documentation libraries using the dev option, i.e., pip install -e '.[dev]'

If you plan to make changes to the code then please make regular use of the following tools to verify the codebase while you work:

• pre-commit: run pre-commit install in your command line to load inbuilt checks that will run every time you commit your changes. The checks are: 1. check no large files have been staged, 2. lint python files for major errors, 3. format python files to conform with the pep8 standard. You can also run these checks yourself at any time to ensure staged changes are clean by simple calling pre-commit.
• pytest - run the unit test suite, check test coverage, and test that the example notebooks successfully run.
• pytest -p memray -m "high_mem" --no-cov (not available on Windows) - after installing memray (mamba install memray pytest-memray), test that memory and time performance does not exceed benchmarks.

For more information, see our documentation.

Building the documentation

If you are unable to access the online documentation, you can build the documentation locally. First, install a development environment of PAM, then deploy the documentation using mike:

mike deploy 0.2
mike serve

Then you can view the documentation in a browser at http://localhost:8000/.