Feature Stores for the 'diseasy' Framework.
diseasystore
Overview
The diseasystore
package provides feature stores implemented in R specifically designed for serve disease data for epidemic preparedness.
What makes a diseasystore
special, is that features can be automatically coupled and stratified within the diseasystore
package. Consult the Quick start vignette to see it in action (vignette("quick-start", package = "diseasystore")
).
The package forms the data-backbone of the {diseasy}
package.
Handling of diverse data sources
Different data sources are handled by individual diseasystores
which each facilitate access to the relevant disease data for the given data source.
Data for different diseases will typically be structured in different ways. The diseasystore
package currently implements the Google Health COVID-19 Open Repository with more diseasystores
on the way.
The diseasystore
package is designed to handle both individual-level data (examples to come) and semi-aggregated (typically publicly available) data.
If the data is at the individual-level, the feature store is fully dynamic and can adapt to (virtually) any stratification that the user specifies. If the data conversely is semi-aggregated, the data can only be stratified at the levels of the semi-aggregation (or at higher levels).
Installation
# Install diseasystore from CRAN:
install.packages("diseasystore")
# Alternatively, install the development version from github:
# install.packages("devtools")
devtools::install_github("ssi-dk/diseasystore")
Usage
For basic usage examples, see vignette("diseasystore")
.