In a previous post we introduced the LinDA project and its motivations in general. Here we focus on the challenges that are associated with working with Linked Data, and on what the LinDA project is going to offer beyond the current state of the art.
The current state of the art of Linked Data tools suffers from the following deficiencies:
- The tools, frameworks, and libraries are often academic, not integrated, and hard to use.
- They often use custom or proprietary means to describe the transformations to be performed; these transformation rules cannot be easily shared between components and frameworks.
- Naive approaches to transform various data input formats are not suitable for certain business needs. E.g., common CSV transformation engines lack support for bidirectional transformations and dynamic binding.
The LinDA project has made it its motto to assist users in Small and Medium Enterprises (SMEs) in this procedure. To this end, LinDA will make contributions to the following fields:
- Open Data Renovation. This refers to transforming data in various formats to RDF, or, perhaps more often, to providing an interface for transforming data to RDF upon request. (Such an interface is called a SPARQL endpoint.) Conversion of relational data (i. e., databases) will have the highest priority.
- Open Data Consumption. The data acquired in the previous step will be processed further to support the organization´s business processes. For example, complex data and their interactions may be visualized, and new relations can be revealed as a result of the interconnections of the data sets.
In both fields, the ecosystems are progressing fast and steadily. The LinDA tool chain focuses on selecting tools from the rich but unstructured supplies, and making them usable for the envisioned target group. The LinDA tool chain will also focus on certain standards, “traditional” and emerging ones, such as the data format RDF and the mapping language R2RML. LinDA will also comprise a Mapping Designer, which will capitalize on R2RML and will allow for domain semantic driven mappings and reuse of mappings.
Transforming data to RDF implies full formalization of previously semi-formalized data. It is therefore a notoriously difficult endeavour, which can be automatized only partially. LinDA supports this procedure by supplying a vocabulary and metadata repository, and puts it into the hands of domain experts rather than computer scientists.