An R package for modeling and forecasting direct-contact and vector-borne infectious diseases. This repo is a fork of predsci/DICE4 and contains the code to reproduce results from our manuscript 'Accurate influenza forecasts using type-specific incidence data for small geographical units'. See 'Data Access' below for data distribution guidelines.
> library(devtools)
> install_git(url="https://github.com/predsci/DICE_quidel_manuscript", subdir="dice")
NOTE: This method is convenient, but it may still be worthwhile to download the repository (see next subsection) to a user directory. The scripts in examples/ directory and the manual (dice/vignettes/dice.pdf) are quite useful.
Navigate to your preferred directory
$ cd mydir
Download the repository from GitHub (requires git command line tools)
$ git clone https://github.com/predsci/DICE_quidel_manuscript.git
Navigate into the local repo directory
$ cd DICE_quidel_manuscript
Use python script to compile from source
$ ./compile.py
NOTE: If you do not wish to or cannot install DICE globally, it can also be installed to a local R-library using 'R CMD build dice' and 'R CMD INSTALL -l /my_lib_loc dice'
If you were referred here from the manuscript 'Accurate influenza forecasts using
type-specific incidence data for small geographical units', see the example script
examples/Quidel_Manuscript.R to reproduce results.
In general, the scripts in examples/ are a good way to get started and the manual
dice/vignettes/dice.pdf contains much more detailed information along with some
walk-throughs. Help pages for DICE and runDICE are also good starting points:
> library(DICE)
> ?DICE
> ?runDICE
Most of the data that DICE works with is publicly available. To get a copy of the aggregated county data from 'Accurate influenza forecasts using type-specific incidence data for small geographical units', email jturtle at predsci.com or s.riley at imperial.ac.uk . In most cases this data is freely available for non-profit pursuits. Please include in your request a statement that the data will be used for non-commercial purposes.