Version 3.1.0 Released 2020 June 26

The Epidemic Prognosis Incorporating Disease and Environmental Monitoring for Integrated Assessment (EPIDEMIA) Forecasting System is a set of tools coded in free, open-access software, that integrate surveillance and environmental data to model and create short-term forecasts for environmentally-mediated diseases. The updated EPIDEMIA forecasting system now is a core set of two packages for supplying functions, epidemiar and clusterapply, and a demonstration data project, epidemiar-demo, for simulated data and example scripts. There is also a fourth package available, epidemia-gee, that demonstrates the ability to connect directly to Google Earth Engine from R via python (with a sample script included in epidemiar-demo).

This is a non-backwards compatible update. Please make sure you have the latest epidemiar-demo (link below) to use with this version of the package.

epidemiar v3.1.0
Expanded model options:
Modeling is no longer limited to two family options (Poisson and negative binomial). Now any model form/family supported by mgcv::bam() can be used.
The user has an option to include a cyclical term in the model to account for seasonality.
The user has an option to interpolate (estimate) any missing epidemiological data.
The user has an option to use the environmental variable data values directly or use the anomalies (difference from expected values) in the modeling. The user can now set a specific forecast start date, irrespective of the date of last known epidemiological data. Previously, the forecast would automatically start the week following the date of the last observed case counts. Now the user can request a specific week to start forecasting. This is independent from user settings on the length of the forecast period (how many weeks in the future/beyond the start week), total report length, or early warning period.
The user can select between two different methods in the modeling of long-term trends and lagged environmental variables: thin plate splines (see clusterapply package below) or modified b-splines.

Revised input parameter scheme:
As epidemiar is designed to be flexible, this therefore leads to a great number of options that can be set. To organize the input parameters to the main function, only absolutely required data and parameters are now top level. All other settings are now found inside the ‘report_settings’ parameter list. All optional settings have defaults, however close attention needs to be paid on if the default makes sense for the respective model and data.

Validation:
The validation output also includes a per week (per geogroup per week-ahead) entry as well for additional investigation and graphs in the formatted validation report of epidemiar-demo.

Robustness and bug fixes:
Removes limitation in epidemiological data set on explicit missing data only - functions can now handle implicit missing data as well.
Multiple bug fixes, including correcting an issue with environmental data gap-filling for 'sum' type variables when the previous week had high values.

We suggest installing using remotes::install_github() function as this will automatically install dependent packages (the modified build_opts should allow vignettes to be installed also).

remotes::install_github("ecograph/[email protected]", build = TRUE, build_opts = c("--no-resave-data", "--no-manual"))

If you are not installing via remotes::install_github(), and you are installing from the binary/source files: make sure that you have all the package dependencies installed (listed in the DESCRIPTION), as this does not automagically happen.

See the included package vignettes, and also epidemiar-demo for a full demo and walkthrough of how to create environmentally-mediated disease forecast reports. Package vignettes are also attached to this release in pdf form.

Other EPIDEMIA system items:
epidemiar: https://github.com/EcoGRAPH/epidemiar/releases/latest
clusterapply: https://github.com/EcoGRAPH/clusterapply/releases/latest
epidemiar-demo: https://github.com/EcoGRAPH/epidemiar-demo/releases/latest
epidemia-gee: https://github.com/EcoGRAPH/epidemia_gee/releases/latest