The Knowledge Representation and Reasoning research group at the VU University Amsterdam has a project, the LOD Laundromat, where a significant subset of the Linked Open Data (LOD) cloud is crawled and converted in a standards-compliant way to remove any inconsistencies such as syntax errors, duplicates and black nodes. So far this resulted in 600.000 knowledge graphs with a combined total of over 40 trillion edges. These graphs are very heterogeneous, running from 10 edges to 100 million edges per graph, build by scripts or by hand, with various topics. All graphs combined are currently assumed to form a complex system.
Visualizing this data on various levels can give insight in what kinds of data is present more clearly and intuitively when done right. There are a number of interesting or challenging aspects to this: how can you show such a large amount of data clearly, will semantic notions be visually recognizable and the other way around, will parts of visualizations result in new semantic notions? Especially the last question can be interesting with a complex system: some properties of the data may only be observed in a different scale of abstraction.

Suzanne Tolmeijer from the Vrije Universiteit Amsterdam will visit Katy Börner’s Lab at Indiana Bloomington.