The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.

arealDB

CRAN_Status_Badge DOI

R-CMD-check Coverage Status Lifecycle:maturing

Overview

Areal data are a rather frequent type of data in many applications of the environmental and socioeconomic sciences, where various aspects are summarized for particular areas such as administrative territories. Many of those applications surpass the spatial, temporal or thematic scope of any single data source, so that data must be harmonised and normalised across many distinct standards. arealDB has been developed for the purpose of building a standardised database encompassing all issues that come with this. In the current, revised version, it makes use of the ontologics R-package to harmonise the names of territories (from geometries) and the target variables (from tables). Moreover, it uses the tabshiftr R-package to reshape disorganised tabular data into a common format.

Schematic overview of the result

Installation

  1. Install the official version from CRAN:
install.packages("arealDB")

or the latest development version from github:

devtools::install_github("luckinet/arealDB")
  1. Read the paper for a more scientific background, or study the vignette on the arealDB pipeline.

Getting started

To study how arealDB works, one can make use of the function makeExampleDB(), where the full process of building an areal database can be “simulated” with dummy data. This can be used to train yourself on a particular step based on a fully valid database up until a certain stage of the process. For instance, to set up database that has merely just been started, but doesn’t contain any thematic data yet, one would use makeExampleDB(path = paste0(tempdir(), "/newDB"), until = "start_arealDB").

In principle, arealDB follows a simple process involving three stages:

  1. Setup the database (stage1): To start a new areal database, one needs to specify a gazetteer that contains the valid names of territories and optionally an ontology that contains the terms that valid for the target variable(s).
  2. Register dataseries, geometries and tables (stage2): Data are typically part of some series of data, a collection of tables that are formatted by the same entity and typically in the same or very similar way across all tables. A data item that shall be inserted into a database is registered by calling the registration function. This function registers the configuration of the file and/or records meta-data. For areal data there is typically a spatial reference to which the data shall be associated. These geometries would be registered before the tables containing the data.
  3. Normalize geometries and tables (stage3): After registering all relevant data, they are reshaped into a standardized database format. In this process terms of territories and target variables are “translated” according to gazetteer and ontology, spatial data are standardized and validated, thematic data are standardized and matched to spatial data, and the spatial data are matched with the optionally already existing spatial database, for instance if that has been built off the GADM or GAUL standardized datasets.

Acknowledgement

This work was supported by funding to Carsten Meyer through the Flexpool mechanism of the German Centre for Integrative Biodiversity Research (iDiv) (FZT-118, DFG).

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.