The Linked Open Data (LOD) Cloud contains more than 31 billion triples. This wealth of knowledge is expressed by using a large number of vocabularies (e.g., FOAF, VCard, SKOS), leading to a schema mismatch between the different knowledge bases that express knowledge about the same types of entities (e.g., persons). Mismatches at instance level are similarly common, including the use of different units of measure (e.g., cm vs. inch for heights) or data types (e.g., strings instead of doubles for geographical coordinates of locations). The integration of this data is yet central for tasks such as large-scale inference, cross-ontology question answering, graph traversal and especially business applications that require all data to be available from their local servers in a particular format.
The aim of the ALOE (Assisted Linked data cOnsumption Engine) project is to provide an engine for the semi-automatic generation of consumption configurations. For this purpose, ALOE can automatically discover class and property mappings across endpoints even when no schema information is available. In addition, ALOE can generate initial specifications for the consumption of Linked Data. For this purpose, it provides several functions for transforming the data from the source knowledge base into a format that corresponds to that of the target knowledge base. Therewith, ALOE enables lay and experienced users to consume Linked Data with great ease. The ALOE process can be controlled via the interface shown below. Please note that the demo only consumes 1000 triples.