Towards a Broader Adoption

Linked Data principles have been introduced into a wide variety of application domains, e.g. publishing statistical data and interpretation of statistics [5], improving tourism experience [6], pharmaceutical R&D data sharing [7], crowdsourcing in emergency management [4], etc. A few years ago, our analysis of the adoption of Semantic Web technologies by enterprises [2] has shown that companies benefit from features that improve data sharing and re-use (57 %), improve searching (57 %), allow incremental modelling (26 %), explicit content relation (24 %), identifying new relationships (17 %), dynamic content generation (14 %), personalization (10 %), open modeling (12 %), rapid response to change (10 %), reducing time to market (5 %), and automation (5 %). Of the features the LOD2 Statistical Workbench provides functionality improving the following areas: data share and re-use, improved search, explicit content relation, identifying new relationships, open model and automation. The LOD2 Statistical Workbench supports both publishers and consumers of Linked Data such as national statistical offices (institutes), national banks, publication offices, etc. Some of those collaborated with us as an early adopter of the approach.

Use Case 1: Digital Agenda Scoreboard

In the course of the LOD2 PUBLINK 2010 activities, the digital agenda scoreboard[1] has been created as the first web portal exploiting the RDF Data Cube. The Digital Agenda Scoreboard provides insight on how 'digital' Europa is. By using an early version of CubeViz the trends are visualized embedded in human readable scenario. Behind the scenes the data is provided and aggregated in a Virtuoso store according to the Data Cube vocabulary. This data is made available to the public in different formats including the RDF representation.

Use Case 2: Statistical Office of the Republic of Serbia (SORS)

In the course of the LOD2 PUBLINK 2011 activities, the SORS public data was integrated into the LOD cloud via the Serbian CKAN [1]. The Serbian CKAN is a metadata repository to be used for dissemination purposes by Serbian national institutions. Maintenance activities include identifying changes in the dissemination data (new public data, changes on metadata level) and fixing the mapping process (from XML to RDF) accordingly. The SORS is in the process of adopting the LOD2 Statistical Workbench[2] that will allow the users to automatically publish data (in the existing and new packages) to the Serbian CKAN.

Use Case 3: Business Registers Agency

In the course of the LOD2 PUBLINK 2012 activities, example data from the Regional Development Measures and Incentives Register was triplified using the LOD2 Statistical Workbench and registered with the Serbian CKAN. The data is reachable via the Serbian CKAN

[3] and can be explored through a prototype application[4].

Challenges Faced by Early Adopters

For each early adopter the publishing of the statistical data as Linked Data has influenced their data publishing process. The Linked Data vision impacts the publication process typically more deeply as it sees data from a more universal perspective and not as an isolated piece of information. When the statistical observations becomes real Linked Data, it means that also the dimensions have to become Linked Data, and this typically means that other organizations that maintain the dimensions have to be consulted.

Therefore in addition to our technological support in the identifying the set of applicable vocabularies and specifying the transformation flow to be setup, there has been an important activity in supporting the early adopters with their relationship with their data suppliers.

Over time the technological support has been improved. Whereas for the first case, the Digital Agenda Scoreboard, many of the transformation steps and data cleaning steps had to be done manually, they are for the more recent applications semi-automated.

Our approach to customize the LOD2 stack not only holds for the statistical domain, but can be applied other domains as well. For instance in the GeoKnow[5] project the GeoKnow Generator is being created for the support of geo-spatial Linked Data.

  • [1]
  • [2]
  • [3]


  • [4]
  • [5]
< Prev   CONTENTS   Next >