Overview

Herein lies the dataset and code used to produce the results presented in Diffusion in Colocation Contact Networks: the Impact of Nodal Spatiotemporal Dynamics. To run all simulations and produce all results one can simply run main.sh.

Execution environment & dependencies

Execution environment

The code was executed on Ubuntu Linux 12.04. Note that in addition to the dependencies outlined below, other unix-based environments may require additional dependencies to be installed.

Dependencies

Here is a list of the dependencies of the main script:

• GNU Parallel
• Go
• Matplotlib
• Numpy
• NetworkX

If you identify further dependencies please let us know.

Code structure & documentation

Structure

There is a single entry point script i.e. main.sh which when executed will run all simulations and produce all numerical outputs and their derivatives (figures, tables, scalar constants etcetera). These outputs are written to the OUT_DIR directory specified in main.sh.

The code relies heavily on the unix pipeline and standard unix commands to manipulate the data. Most of the more complex computations are written as Python scripts or Go scripts which accept input from stdin and either write output to stdout or directly to a file if the output is a figure.

Documentation

Extensive comments can be found throughout the code base.

Dataset

There are two .csv files provided in this distribution: (i) uq.csv and (ii) st_lucia.csv. Both files consist of a set of lines of the form <mac, start, end, ap> describing a mac having a session between start and end at ap. st_lucia.csv is the dataset for only the UQ St Lucia campus. uq.csv is a superset dataset for all UQ-affiliated sites. These two files consitute the minimal dataset necessary to reproduce the results presented in the paper. Note that mac and ap have both been minified i.e. the original identities have been remapped to a set of more compact identifiers.