Simulation of an epidemic in a heterogeneous agent population to explore the effects of different mitigative measures such as recommendations, policies as well as individual mitigative behaviors such as mask wearing.
This model was build in Julia 1.4.2. It depends on the third-party Julia packages Agents.jl, DataFrames.jl, LightGraphs.jl, StatsBase.jl, Feather.jl, JSON.jl, Distributions.jl, GraphIO.jl, Pipe.jl, Query.jl and ProgressMeter.jl which can be installed directly from within the Julia package manager Pkg.jl. To switch to package manager, enter ] into the REPL. To install packages, enter the following code (ommitting the .jl):
(@v1.4) pkg> add <PACKAGE_NAME>
A further requirement is the file archiver 7-zip. 7-zip needs to be in the local PATH variable. To test whether 7-zip is properly setup, run the command 7z in a terminal.
You can test if the code runs by executing the test script:
julia>include("test_module.jl")
To reproduce my Master thesis experiments, execute the following:
julia>include("experiments.jl")
This will run all simulations that are setup in the scripts in the config folder. Please be advised to only run this script if you have enough storage on your device. Even though the data is compressed by the module, the simulations create a substantial amount of data and run for an extended period of time.
Once you have run the simulations, you can reproduce all analyses that I conducted. The analysis scripts for all experiments can be found in the folder analysis. Each subfolder contains an .Rmd script that recreates all analyses that have been run on the simulation data. The analyses scripts were written under R 4.0.2 and require the following R packages: tidyverse, magrittr, ggpubr, igraph, arrow and viridis. All of these packages are available from CRAN:
R>install.packages("<PACKAGE_NAME>")
The data can also be found at https://osf.io/kfmsx/.