Based on the Open Data Charter of G8 in 2013, the Japanese government published a dataset covering approximately 4.4 million corporations in December 2015, but the dataset is merely rated as 3-star in the 5-star rating system of Open Data. To improve usability, we de ned a schema for corporation data, converted the dataset into 5-star using this schema, and published the created dataset under Creative Commons Attribution 4.0 License. As far as we know, eight datasets currently refer to ours, which makes the degree of 5-star stronger. To demonstrate our dataset, we applied this dataset to a visualization system showing various information of a corporation.
The G8 leaders signed a charter on Open Data on June 18, 2013 [
Based on this background, the governments in the world aim to publish
corporation datasets. Meanwhile, according to a survey of LOD datasets, no
dataset of corporations is included in the highest referred datasets [
The objective of our work is to build a 5-star Japanese corporation register dataset. In Japan, the National Tax Agency (NTA) has published a dataset covering approximately 4.4 million corporations on December 01, 2015, but the dataset is merely rated as 3-star in the 5-star rating system. Our policy, which we believe is also common in the LOD community, is that low-star datasets must be converted into 5-star as early as possible for strengthening the power of LOD. Based on this policy, we de ned a schema for corporation data, converted the Japanese dataset into RDF using this schema, appended links between our dataset and other data, and published the rst version of this dataset under Creative Commons Attribution 4.0 License on December 09, 20151, only eight days after the publication date of the original dataset. As far as we know, eight datasets currently refer to ours, which makes the degree of 5-star stronger. We 1 http://lod4all.net/datasetdetail.html?graph=http://lod4all.net/graph/ corporate also applied this enriched data to a visualization system for browsing a corporation.
The following sections describe our proposed schema design of the dataset, the building process and the demonstration. We nish the paper with future directions.
Tim Berners-Lee suggested a 5-star deployment scheme for Linked Data [
First, we de ne the URIs of corporations that are independent of properties
of a corporation and are not affected by changes in properties [
Next, we investigated the schema design [
We built an RDF-modeled dataset from the NTA dataset using the URIs and schema design. Additionally, we added a link to the corresponding URI in DBpedia3 with foaf:primaryTopic, and attached the corresponding ticker symbol and EDINET code4 as literals with skos:notation using a simple string matching method. We also added a geographical point of an address as literals with geo:lat and geo:long using a geocoding method.
The following example5 shows a corporation in our dataset in Terse RDF Triple Language. <https:k/o/sl:opdr4eaflLla.bneelt/"rFeusjoiutrscue/Lciomriptoerda"t@ee/njp;/1020001071491> a org:Organization ; org:identifier "1020001071491"^^<http://lod4all.net/ontology/corporate/jp/corporateNumber> ; # Corporation identifier of NTA skos:notation "6702"^^<http://lod4all.net/ontology/corporate/jp/stockTickerNumber> ; # Ticker symbol of Tokyo Stock Exchange skos:notation "E01766"^^<http://lod4all.net/ontology/corporate/jp/edinetNumber> ; # EDINET code fddoccattf:::ispsursbiujmeeadcrty"T2<o0hp1ti5t-cp1:0/<-/h0lt5otTdp:04/0a:l/j0l0.a.:n0edb0t"/pe^r^edixssoa.du:rordcaeg/t/ecreToirsompeourrac;ete//jp/co>nc;ep#t#Fujitsu > ; # Stock corporation dct:modified "2015-10-15T00:00:00"^^xsd:dateTime ; org:hasRegisteredSite [ a org:Site ; org:siteAddress [ schema:postalCode "2110053" ; schema:addressCountry "Japan"@en ; gn:countryCode "JP" ; schema:addressRegion "Kanagawa"@en ; 2 ISO 3166-1 alpha-2 country code: http://www.iso.org/iso/country_codes 3 http://dbpedia.org/ 4 EDINET (http://disclosure.edinet-fsa.go.jp/EKW0EZ1001.html) is a system of Financial Services Agency of Japan for Annual Securities Report, and an EDINET code is an identi er of a corporation. 5 Literals in the example are expressed in English with @en for understandability.
They are actually expressed in Japanese with @ja. ] ; ] ; org:changedBy [ adcotr:gs:uCbhjaencgteElv4ean-tco;rporate-jp-concept: ; # Absorption-type merger. prov:atTime "2015-10-15T00:00:00"^^xsd:dateTime ; dct:description "On October 1, 2015,
Fujitsu Wireless Systems Limited (9030001085368) located at Nakasone 1376, Kumagaya, Saitama was merged."@en ; ] .
This example contains the data that are included in the dataset of NTA; a name (skos:prefLabel), an address (org:hasRegisteredSite), and an identier (org:identifier). Additionally, information about important events such as M&A, and meta information about data update itself such as change of address (org:changedBy) is also included.
On December 09, 2015, we released the dataset on the SPARQL endpoint
and the content negotiation framework in LOD4ALL6 [
To con rm utilization of our dataset by other people, we counted links from other datasets on LOD4ALL to ours. We found that as much as eight datasets currently link to our dataset.
We present a demonstration system for con rming the usefulness of our published dataset. The system inputs an entity from a user and displays an HTML page with multiple components, each of which is constructed with data linked to the entity.
Fig. 1 shows the snapshot for the corporation, Fujitsu Limited. If only the NTA-origin data were available, the snapshot would contain only the box denoted as NTA. By virtue of the enriched data we created, it contains various information as shown. For example, the DBpedia data obtained with links provides the logo in the upper left and the abstract in the upper right. Each component in the gure is generated by a JavaScript program with a function of querying a SPARQL endpoint. Below is the example of a SPARQL query used in the component showing the logo:
SELEC<Tht?tipm:g//WlHoEdR4Eal{l.net/resource/corporate/jp/1020001071491> <http://xmlns.com/foaf/0.1/primaryTopic> ?dbpedia . } ?dbpedia <http://dbpedia.org/ontology/thumbnail> ?img .
We de ned a schema design of our Linked Data of corporations and published the dataset for about 4.4 million corporations by Creative Commons Attribution 4.0 License. Our dataset has already been linked from eight datasets, which indicates that our dataset contributes to enhancement of the LOD cloud. We also demonstrated our dataset using our visualization system that aggregates related data from various sources. 6 http://lod4all.net/
DBpedia
NTA EDINET (3) (4) (5) (6) (7)
Our future direction is enhancement of the value of our dataset. First, we will build additional LOD datasets of corporation register data from other countries. Secondly, we will append extra relationships to several other datasets. These expansions will contribute to the growth of data associated with corporations in LOD.