The overall application is a tool to help developers navigate, understand and make use of the large amount of data generated from testing runs during product development. It's all rather proprietary, and I've scrubbed out any context that might make it understandable. What's left is a directed graph of nodes, each with 16 or so numeric attributes. Behind the scenes it uses a force directed simulation to lay out the nodes and links.
Hover over nodes to see their immediate links, click to see attributes and additional per-node controls, double click to <redacted>. Click on "Change Scale" to scale size & color by different attributes. Adjust max depth to control the number of hops from the "main" nodes displayed. Below the table is a selector to view different data sets (not a production feature, so bugs may intrude).
This visualization uses the following libraries: