Browsing Linked Data Forest
Browsing Linked Data Forest
by Paolo Nesi, University of Florence. — December 2014
I started several years ago on parallel architectures for semantic computing, mainly for addressing back office problems behind large media companies as broadcasters and social networks. Thus, in practice media transcoding for content ingestion and adaptation, content indexing for querying, production of recommendations, content protection and licensing, and user profiling. A change of pass and scale arrived to constraining to change paradigm for addressing the same problems in same sense at larger scale. Thus, almost all these problems have been identified to be as big data problems.
Recently, I moved on smart city issues addressing open data and private data interoperability. The issues and problems of the past regarding lack of quality and interoperability among metadata archives now are appearing regarding open data and private data generated from multiple databases, public administrations, standards, people and moods. Thus, i realized the need of developing a new generation of data mining tools and solutions to generate semantic interoperable RDF stores, again for indexing, reasoning and recommendations.
In the present generation of smart city problems, the main challenge is probably the identification of what can be done with the accessible data, and may be which data are missing to infer, predict or deduce some effect or results on the city. In most cases, the idea and the city expectations are in the mind of city strategists which are far from data world and data engineering issues. This approach constrained me, and my lab, to continuously experiment on adding new data and changing approaches, to produce new generation of solutions to make complex data model understandable.
With this aim we are developing a number of algorithms and tools for OD/LOD data ingestion, mining and reconciliation; and also for multi sparql end point LOD analysis and browsing, and for decision making. So far, first results are accessible based on data about Florence area which is one the most promising smart city in Italy.
Among the tools we developed, we think that, one of the most useful and promising tool for developers and for analysis is the Linked Open Graph.
It presently allows you browsing, querying and studying graphs involving multiple sparql end points to create your own view on top of multiple RDF stored (e.g., Europeana, Getty voc, LinkedGeoData, etc., and your own “private” RDF stores and in your smart city and cloud as we have). With LOG.DISIT.org, you can save the analysis, and share the graph with your colleagues via email, and may be embed the view on your web pages and services, re-edit them again for further analysis and update, etc. An open access document is available on http://www.sciencedirect.com/science/article/pii/S1045926X14000962