Repository for the FoodWeb hackathon graph experiment

Given:

• A complex graph describing a relatively simple underlying path with abundant variations, with a matrix representation of its connectivity (i.e., an edge matrix) We can:
• Reorder the graph to “most forward-travelling order”
  • In matrix form, we are minimizing the sum of the elements under the diagonal
• Group elements simply
  • To group nodes 2 and 3, take the edge matrix, take row 2 and row 3, combine by summing corresponding elements in each; similarly with columns 2 and 3 (order of these two operations is irrelevant This means: If we have an "expected path":
• Nodes between those in the “expected path,” when in ideal order, often should be groupable based simply on their ordering
• We think this is the case, but hasn’t been demonstrated yet
• Would like to demonstrate dynamically

• Put together a dynamic graph visualization that lets one take a large, complex graph, and allow one to:
  • Put the graph in optimal order
  • Dynamically change graph order
  • Dynamically group/ungroup elements in the graph
    • Maybe automatically with “core path” specification With all changes reflected in the graph vis
• Ideal comparison:
  • Force-directed vis vs
  • Ordered vis vs
  • Ordered and aggregated vis

• Install Node
• Run setup.sh
• Run yarn build

• Run go run server.go
• Connect to [http://localhost:6789]

• Run build.sh to build a Docker container
• Run run.sh to run the Docker container
• Connect to [http://localhost:6789]