Chapter 7. How to organize a chart
This chapter covers
- Layouts and why they’re important
- The force-directed layout variants and when they’re helpful
- Other useful layouts for specific data structures
- The drawbacks of 3D rendering
Unlike many other data visualizations, in a graph visualization the actual location of nodes on the screen has no inherent meaning. Nothing is different about a node that’s located at the upper-left corner of the screen versus one at the bottom right. This is in stark contrast to something like an XY scatter plot where the location of the item on a Cartesian plane tells you the value of specific properties of that item. In node-link-style graph visualizations, the location of the node is determined by convenience and readability. Nodes should be placed in locations that make the chart less cluttered and easier to read. Unfortunately, automated layout algorithms aren’t magic. Once the number of nodes and edges on a chart gets into the hundreds, no layout will manage to make it readable. One of the most common requests I get with larger charts is to detangle the links, or prevent links from crossing one another. With most charts, this isn’t possible because of the limitations of 2D geometry.