0.1.8 regions: working with sub-national statistics. Performance improvement

This is minor fix to a performance issue with the recently released version 0.1.7.

• Removing a bottleneck from validate_geo_code(). The NUTS exceptions are essentially constants, and it was unnecessary to calculate them each time these functions were running. They were moved to the nuts_exceptions dataset.
• Citation info was given the the vignette articles.
• Some error messages in assertions were made clearer are more consistent.

Minor release with new visualization vignette article

This is a minor release with plenty of improvements in the code, the unit tests and the documentation. The package is now working with rlang and tidyverse 1.0+, and uses more meaningful and descriptive assertions and error or warning messages.

A new vignette shows how to place the problem areas on a map, and how to draw the map.

Second CRAN release

• Documentation improvements.
• validate_country_nuts_countries is now follows dplyr 1.0, this makes the code more readable.
• validate_nuts_regions is validating non-EU NUTS-like regions as valid if they will be added to NUTS2021. These regional codes, while legally not part of the NUTS2016 typology, are valid and can be placed on the maps created by EuroGeographics maps provided by Eurostat.
  Italy, Portugal, the United Kingdom, Estonia, Slovenia,Latvia pseudo-NUTS3 codes in google_nuts_matchtable.
• New correspondence table for converstion between Local Administration Units (LAUs) and NUTS within the European Union and some other European countries.
• Correction of a small bug with input data frames that already have columns that the validation result should contain.
• Wordlist added for spell checking with geographical exceptions.

regions 0.1.3 is on CRAN

regions is released on CRAN. This package helps creating data time series and data panels for sub-national, i.e. regional, provincial, federal state, county and other lower level data.

While we often take the stability of national boundaries for granted, just within the EU there are almost 2000 regional boundaries, and dozens change every three years. Globally, there are tens of thousands of provinces, states, counties, regions, parishes and other areas that may be used for statistical purposes, with hundreds of changes annually. Joining data from Google and Eurostat, from U.S. statistics and OECD is an absolute nightmare, partly because the ISO standard for lower level administration is extremely ill-suited for statistical applications due to its poor metadata structure.

For European data tables made since 1999, not so much anymore. We have encoded thousands of European region versions, and hundreds used by Google to allow the creation of far more detailed statistical analysis that nationally aggregated data would allow.

First CRAN release candidate

The first CRAN release 0.1.0

Regions is an offspring of the eurostat package. It started as a patch to the problematic transition of Eurostat’s regional statistics from the NUTS2013 to the NUTS2016 regional typology, which made it impossible to create time series, data panels or data joins among different data tables. Eurostat’s regional statistics is not tidy, because in the ‘geo’ ID it mixes together with incompatible identifiers from the NUTS2016 (current base case), and the NUTS2010 and often earlier typologies.

Current features:

an extended validation of country codes and European NUTS regional coding starting from 1999
recoding and relabelling across European NUTS typologies, and general functions that can work with any national, ISO 3166-2 or OECD typology as inputs.
projecting data from larger sub-national divisions to smaller ones
two vignettes explaining this functionality.
Planned features:

Incorporation of new typologies, such as OECD's, Google's, and other major data sources.
Better coordination between rOpenGov packages, particularly eurostat and eurostat_geodata.
New imputation functionality based on re-aggregation, re-weighting.
Vignette on joining data from different typologies, such as Google Trends data with Eurostat data.
The main repo is rOpenGov/regions, where you can send pull requests to the devel branch. You can directly collaborate with the package author on antaldaniel/regions, too, but the core development will hopefully shift to rOpenGov.

First CRAN release

The first CRAN release 0.1.0

Regions is an offspring of the eurostat package. It started as a patch to the problematic transition of Eurostat’s regional statistics from the NUTS2013 to the NUTS2016 regional typology, which made it impossible to create time series, data panels or data joins among different data tables. Eurostat’s regional statistics is not tidy, because in the ‘geo’ ID it mixes together with incompatible identifiers from the NUTS2016 (current base case), and the NUTS2010 and often earlier typologies.

Current features:

• an extended validation of country codes and European NUTS regional coding starting from 1999
• recoding and relabelling across European NUTS typologies, and general functions that can work with any national, ISO 3166-2 or OECD typology as inputs.
• projecting data from larger sub-national divisions to smaller ones
• two vignettes explaining this functionality.

Planned features:

• Incorporation of new typologies, such as OECD's, Google's, and other major data sources.
• Better coordination between rOpenGov packages, particularly eurostat and eurostat_geodata.
• New imputation functionality based on re-aggregation, re-weighting.
• Vignette on joining data from different typologies, such as Google Trends data with Eurostat data.

The main repo is rOpenGov/regions, where you can send pull requests to the devel branch. You can directly collaborate with the package author on antaldaniel/regions, too, but the core development will hopefully shift to rOpenGov.