*
We are delighted to announce the first release of the LODVader v1.0
available at <http://lodvader.aksw.org/>http://lodvader.aksw.org/driven
by requirements developed in the LIDER project
<http://lider-project.eu/>http://lider-project.eu/.
What is LODVader?
LODVader stands for LOD Visualization, Analytics and DiscovEry in
Real-time, and is available as a REST API. LODVader indexes RDF datasets
fetching statistical data for analysis and creating a diagram allowing
users to visualize links among datasets.
How does LODVader works?
LODVader parses your dataset description file that might be in different
formats such as VoID, DCAT and DataID. Then, we stream your RDF data in
order to extract links and statistical data, and compare with different
Bloom filters which contains index from other datasets.
The following features are available in the v1.0.
*
Visualization: LODVader supports a multi-layer graph visualization
interface which visualizes datasets and their respective relations.
Moreover, in many cases it is important to identify dataset links
which are not connected within the imported dataset cloud. Therefore
we introduce the novel notion of the Dark Cloud. Using the above
features makes it possible to create a new LOD diagram showing
broken links between source and destination datasets. The broken
links discovery also relies on BFs.
*
Dataset comparison: estimate similarity among different datasets and
perform datasets comparison based on their similarity. Similarity
metrics, such as the Jaccard similarity coefficient, are being applied.
*
Analysis via RDF Streaming: LODVader supports the ability to deal
with RDF streams, which enables the support of different kinds of
RDF input sources. Example RDF data sources are RDF dump files,
SPARQL endpoints or other RDF data streams.
*
Link Extraction: LODVader uses an advanced approach to detect and
extract links between datasets using Bloom filters (BF). The
extraction if performed on-the-fly, when the datasets are being
streamed.
*
Top-N Analysis: Based on the data which was collected during the RDF
streaming process, statistical analysis of each dataset regarding
the top-N used properties, links, relations and similarities are
performed and made available for further use.
*
Dataset Search Index: Based on the BF search index is created by
indexing subjects and objects as BF vectors, thus allowing fast
access to this data for comparison and search operations. In
addition, LODVader allows to search and filter datasets or
ontologies by subject, property and objects.
*
Dataset Statistics: Due to the vast amount of data which is stored
in each dataset, it is important to collect statistical information.
Accurate statistical analysis of each dataset regarding the top-N
used properties, links, relations and similarities is performed and
made available for further analysis.
Moreover, you can check the Wiki, and try our online demo at:
<http://lodvader.aksw.org/>http://lodvader.aksw.org/
Your feedback is more than welcome,
Ciro Baron Neto.
Acknowledgments
*
LODVader is an open-source project maintained by the KILT subgroup
of AKSW at Leipzig University. You can download and deploy the
project from our source code available at GitHub
<https://github.com/AKSW/LODVader>.
*
Special thanks goes to the LODVader team Kay Müller, Martin Brümmer,
Dimitris Kontokostas, Sebastian Hellmann, Diego Esteves.
*
This research activity was funded by grants from the FP7 & H2020 EU
projects ALIGNED (GA~644055), and LIDER (GA-610782) and the CAPES
foundation (Ministry of Education of Brazil) for the given scholarship.
*
