Using Linked Data to facilitate translations between product and industry classifications

Data that may be of use for Industrial Ecology or Industrial Symbiosis often employs a myriad of classification systems describing different industrial sectors or product categories. The provision of concordances can allow... [ view full abstract ]

Data that may be of use for Industrial Ecology or Industrial Symbiosis often employs a myriad of classification systems describing different industrial sectors or product categories. The provision of concordances can allow for translation between these different systems, although this process can be very tedious when translating a large number of classes, or when no direct translation is specified but translation through an intermediate classification system may be needed first.

To address this, we have created a project that uses Linked Data standards to structure existing classification systems and the concordances between them.  There are several reasons why this is advantageous.  First, it allows us to represent the nature of concordance relations between classifications, which indicate if the mappings between codes may be one-to-one, many-to-one, many-to-many, etc.  Secondly, the query language for Linked Data (SPARQL) can be used in creating queries to automatically find translations, even if multiple translations through arbitrary numbers of other intermediate classifications are needed first.  For example, we are not aware of concordances between Combined Nomenclature (CN) and NACE, although it is possible to link these by first translating from CN to SITC to HS to ISIC and finally to NACE.

The larger aim is that this effort is meant to facilitate the process of interlinking diverse datasets which employ these classifications.  For example, point-source pollution databases may document industrial classification codes of facilities along with coordinates, thus giving a picture of industrial geography.  However, for Industrial Symbiosis, it is interesting to also know about the material flows associated with facilities in a region, and this may be deduced by integrating information from Life Cycle Inventory databases which employ another classification.  The system we have created allows for automatically making these connections, which can assist the larger community with better leveraging existing data to gain new insights.