How can users – especially those who don’t have deep roots within the semantic web community – make Linked Data useful to them? It’s not always apparent, says Roberto Garcia, a mind behind the Rhizomik initiative that has produced a tool called Rhizomer. Its approach is to take advantage of the structures that organize data (schemas, thesaurus, ontologies, and so on) and use them to drive the automatic generation of user interfaces tailored to each semantic data-set to be explored.
A project led by members of the GRIHO (Human-Computer Interaction and data integration) research group that is assigned to the Computer Science and Industrial Engineering Department of the University of Lleida, where Garcia is associate professor, the initiative also has led to projects including ReDeFer, a set of tools to move data in an out of the Semantic Web, and various ontologies for multimedia, e-business and news. As for Rhizomer, it accommodates publishing and exploration of Linked Data, with data-set exploration helped by features including an overview to get a full picture of the data-set at hand; zooming and filtering to zoom in on items of interest and filter out uninteresting items; details to arrive at concrete resources of interest; and visualizations tailored to the kind of resource at hand, as the site explains.
In other words, its features are “organized so they support the typical data analysis tasks,” he says. “We are more a contributor from the user perspective of how you interact with that data.”
For example, components like navigation bars are generated automatically from the ontologies related to the data-set under consideration. Some adjustments have been made to traditional information architecture components, he notes, to allow extracting the full potential from highly structured data. For instance, facets can be used to filter but also to pivot between sets of connected resources, like going from the actors born in Canada in its sample Rhizomer LinkedMDB (a Linked Data version of the Internet Movie Database) to the films in which they have participated, and then continue filtering them by those filmed in Canada. “This kind of interaction makes it possible for lay users to build complex semantic queries, which are rendered as sentences to make them aware of what they are getting — for instance “Showing Films where country is Canada and have actor where born is Canada,” he says.
Rhizomer isn’t new but more recently the team behind it has started to go deeper into the user interaction aspect, he says. “There’s a higher-level focus on making data more accessible through the UI.” Semantic data is quite rich, he says, which opens the door to many useful ways to interact with it, but also to difficulties in understanding how to query it and gain value from it.
Another example of Rhizomer on top of different data sets is Rhizomer DBpedia, which Marco Neumann, chairman of the board of directors of the Lotico semantic web meetup community, recently tweeted as being “a bit slow but very, very interesting if you have to work with DBpedia data and discover Semantic Web data.” Garcia recently presented Rhizomer at a Lotico San Francisco meetup.
As one example of how Rhizomer DBpedia can be useful to those who are unfamiliar with the underlying technologies or any of the terms used in its data-set, Garcia asks to imagine a developer who wants to create a mobile app for bird-watching. “One of the difficulties is that using SPARQL for DBpedia means you have to know the specific name they gave a species of bird, the particular names they have given to classes. Or, even more difficult, for properties. In our approach, you see them as you interact with the tool,” he says. Its filter capabilities, for instance, lets users click on common values and then have a list of those features most appropriate to that family of bird to pop up.
There’s also a version demoing a Subset of the NASA Space Flight and Astronaut Dataset:
“The idea is there is a tool you can deploy on any data set or any semantic data based on RDF,” Garcia says. “It will try to generate automatically the different interfaces that let you explore that data set.” Rhizomer also has been deployed in ongoing research and development projects that apply semantic seb technologies to domains like copyright management, XBRL financial data or e-Government. Currently, for example, a local administration in a city in Spain is interested in exploring how the tool can be used with the 5-star Linked Data set it’s developing, Garcia says.
Rhizomer is available as an open Google code project here.