The World Bank is an international development organization that provides access to a comprehensive set of data about countries around the globe. The publicly available statistical data is collected from officially-recognized international sources, and consists of a wide array of observations.
The World Bank Linked Dataspace was created by Sarven Capadisli at LOD2 partner Linked Data Research Centre, DERI. It is one of the case studies in Statistical Linked Dataspaces.
The original datasets: World Bank Climate Change, World Bank Finances, World Bank Projects and Operations, and World Development Indicators were retrieved in XML format via World Bank’s APIs, transformed to RDF/XML and published using the Linked Data principles. Linux shell scripts were written to retrieve the source datasets and their metadata, to transform the data via custom XSLT stylesheet per dataset and metadata into RDF, and to finally load into a Jena TDB RDF store. The Linked Data Pages framework was used to publish the data in both human readable HTML documents and in several RDF formats using content negotiation. Apache Jena’s Fuseki service is used to retrieve the data from the RDF store and provide a publicly available SPARQL endpoint.
Vocabulary terms were extracted from the original World Bank data, hence no noteworthy proprietary terms were created. As bulk of the data was composed of statistical observations, the RDF Data Cube vocabulary was used to represent multi-dimensional data. The SKOS vocabulary was used to create code lists for dimension identifiers in the Data Cubes; SDMX for the statistical information model; the Dublic Core Terms vocabulary for general purpose metadata relations; the VoID vocabulary to provide metadata on datasets.
There is a VoID file which contains metadata for the datasets. The information included, but not limited to is: locations of RDF datadumps, named graphs that are used in the SPARQL endpoint, vocabularies used, dataset size. Statistics for the VoID file is generated using LODStats. The data dumps are available either as individual RDF/XML files or in compressed gzip format.
The dataset contains nearly 170 million RDF triples.
The dataset is interlinked using LIMES (LInk discovery framework for MEtric Spaces) for countries and currencies to DBpedia and Eurostat datasets.
Some of the example use of the data include: visualizations to show charts for development indicators, LodLive as one of the example datasets for graph visualization and navigation.
In addition to adhering to World Bank’s terms of use, the RDF data that is published is licensed under CC0 1.0 Universal Public Domain Dedication. The code which retrieves the World Bank data, transforms it to RDF serializations, and imports to TDB Triple Store uses Apache License 2.0 and is available in Github: csarven/worldbank-linkeddata.
The world-bank-linked-data dataset is registered in the Data Hub and added to the lodcloud group.
At this time, the Linked Data release of World Bank is neither officially approved or endorsed by the World Bank. The effort is, however, informally encouraged and well received by some of its members and surrounding community.
