ZAME: Interactive Large-Scale Graph Visualization

Abstract

INRIA Figure 1: A protein-protein interaction dataset (100,000 nodes and 1,000,000 edges) visualized using ZAME at two different levels of zoom. ABSTRACTWe present the Zoomable Adjacency Matrix Explorer (ZAME), a visualization tool for exploring graphs at a scale of millions of nodes and edges. ZAME is based on an adjacency matrix graph representation aggregated at multiple scales. It allows analysts to explore a graph at many levels, zooming and panning with interactive performance from an overview to the most detailed views. Several components work together in the ZAME tool to make this possible. Efficient matrix ordering algorithms group related elements. Individual data cases are aggregated into higher-order metarepresentations. Aggregates are arranged into a pyramid hierarchy that allows for on-demand paging to GPU shader programs to support smooth multiscale browsing. Using ZAME, we are able to explore the entire French Wikipedia-over 500,000 articles and 6,000,000 links-with interactive performance on standard consumer-level computer hardware.

Honeycomb: Visual Analysis of Large Scale Social Networks

“…This can then be repeated at multiple levels of scale, depending on the size of the input graph. This hierarchical grouping can then be used to create a single interactive visualization that allows the user to browse the network at different levels of scale [18,1,6]. The major problem with these prior approaches is that, without methods of automatically highlighting potentially interesting or anomalous data points, the user can spend a significant amount of time browsing these representations at multiple levels of scale without learning anything new.…”

“…Schematic representation of the network (grey nodes and curved connections) and the aggregation hierarchy. By numbering the nodes in the aggregation hierarchy in a depth first manner (dotted line) and keeping track of the minimum and maximum values encountered during this traversal we can determine the number of edges connecting groups (6)(7)(8) and (9)(10)(11)(12) by running the query: SELECT COUNT * FROM EDGES WHERE 6 'From' 8 AND 9 'To' 12;…”

“…Alternatively, a solution where all edge counts are precomputed and stored in a disk based index can be used [6], although this makes it harder to switch between multiple hierarchies.…”

“…This allows us to interpret connections between higher level aggregations directly, because they relate to a concrete and meaningful grouping. Previous approaches [1,6] have used network clustering or matching algorithms to automatically generate a hierarchy on a network when none was available. Although these approaches offer the similar scalability and corresponding interaction model, interpretation of the resulting visualization is often problematic because it is much harder to put a meaningful label on an individual cluster.…”

Visual Analysis of One-to-Many Matched Graphs

“…The visual analysis of complex data sets is one of the most natural applications of graph drawing technologies (see, e.g., [11,15,20,24,25] for some recent works). A typical application scenario consists of a set of data (nodes) and one or more relationships among these data (each relationship is a set of edges); therefore one is given one or more graphs on the same set of nodes.…”