RelFinder is part of the Visual DataWeb family. The tool enables users to track relationships between given objects in the interactive visualization. RelFinder supports any RDF datasets, provides SPARQL access and operates within open source framework Adobe Flex. The user audience of the tool is quite variable starting from data analysts in such humanities spheres like history, language studies as well as non-humanities areas like biomedical research and others. The tool is web – based, hosted on Google Code and available under GNU General Public License.     

Like any other tool of Visual DataWeb set RelFinder works with minimal technical requirements and offers intuitive user interface. The integrated datasets of the tool include DBpedia, LinkedMDB, Linked Movie database, Linking Open Data (LOD) and Semantic Web Dog Food. All of them can be easily found and linked to through the user interface in the Settings. It’s good to note that successful running of RelFinder, especially if you want to extract data from DBpedia requires Flash Player to be installed in your browser. The tool can be really helpful for those working with knowledge bases and willing to understand the relationships between any objects of interest. A good point in favor of RelFinder that you can make input of any kind of object you are interested in. For example I made a query of Alberta and British Columbia. Visualization of the relationships between objects works pretty smoothly and quickly. A very useful feature of the user interface is its presentation area. The biggest part of the interface is devoted to the visualization representation for the purpose of better readability. According to the developers there are certain principles put into visualization mechanisms to make it specially user – friendly. The subject – predicate – object schema allows user to see the relationships of the objects on the micro and macro levels simultaneously. User – given objects serve as starting nodes in the graph, the links between the objects further turn into labeled and directed edges. Pointing at any particular word on the graph lets user analyze details of any single object while the sidebar lets user work with more generic picture of the connection. In my visualization of Alberta/British Columbia one can see the connectivity level 2 in the sidebar. Other information provided is the link and class categories of the relationships. For example many people in Alberta and BC are connected by their places of birth or involvement in any particular organization, e.g. OfficeHolder or IceHockeyPlayer.

Despite user – friendliness of the tool there can be some errors in the work of the program as RelFinder is still a research prototype - the developers warn. In my view the absence of an option to save the visualization as a file is a light drawback of the tool. Although the possibility to save the link is present and seriously saves time in case you want to get back to your visualization in the future. Visualization features are very strong. One more good feature of RelFinder is its rich selection of languages in its user interface. You can choose the language for the objects and its description in the sidebar, then the dataset will turn on the description or an object in your desired language.  

RelFinder was mentioned in this research abstract available in XML format https://raw.githubusercontent.com/lehkost/ToolXtractor/master/data/xml/2019/HYV_NEN_Eero_Relational_Search_in_Cultural_Heritage_Linked_D.xml