I added a new attribute to the dataset which is region.
In the node-link diagram, states in the same region likely stick together as they have links in between.
States in region which has less number of links spread out in the diagram. For example: look at the light-green color (South Atlantic), Florida is far apart with those states (DC, Maryland, Virginia, West Viginia, Delaware).
Thus, although the dataset is non-spatial, users can explore the relationship of nodes by looking the distance or number of hops in between.
I want to encode the region information into the adjacency matrix view.
Using the same dataset with region attribute, it is impossible (or hard) to encode the region information for a pair that each state in that pair has different region (those in blue color).
Or using two half trianges in each link with region colors?
When ordering by region we can see those states in same group are closer together.
Each row or column has at most two colors. For those have one color (Utah, Michigan, or New Hampshire) that mean they does not have border with state that in different region.
forked from anonymous's block: VI7: Node-link diagram and Adjacency matrix
Modified http://d3js.org/d3.v3.min.js to a secure url
Modified http://d3js.org/colorbrewer.v1.min.js to a secure url
https://d3js.org/d3.v3.min.js
https://d3js.org/colorbrewer.v1.min.js