I came at this project with the goal of combining these three projects into a responsive binary tree that blended three channels of CMYK and spawned new numbers as they touched other nodes. I did none of this.
Having very little interaction with p5 since ICM, I really considered converting the binary tree into Processing where I felt more comfortable, but decided it was more valuable in the long run stumbling through. So I stumbled away.
The above sketch takes the entire text from On the Origin of Species and rates sentences from the corpus that compete for survival. The survival parameters are based on the golden ratio (1: 1.61...) - an ideal score is given to word pairings where the lengths of the first and second words match this ratio. Sentences with a higher overall fitness score appear as larger orbs and have a better chance of meeting other sentences and propagating.
My main roadblock was trying to coax any interactivity from the nodes. This was 90% of my time. I rewrote functions - trying every combination of passing arguments into them with little to show for it. I was able to interact with a single node, but I realized the entire tree was being drawn node by node behind it.
Before completely abandoning this approach, I tried to create three identical binary trees (that were responsive) and separate them by channels - in effect making the CMYK channels on the tree itself rather than the nodes. Although the base node accepted a cyan color, I found the entire tree was very determined to stay in grey-scale.
So at the end of one very long failed avenue, I at least played with the design of the binary tree and made a fractal inspired mirror binary tree.