I've developed some programs with my collaborators for
experimenting with graphs. These program allow one to input a graph (or set of graphs) or enter
parameters for the creation of a random graph; then it tests the graphs for each of these properties:
CFS checks
for having the "constructed from squares" property.
LGT checks whether or not a graph
yields a relatively hyperbolic right-angled coxeter group.
TEA, was written by Robbie Lyman, and computes upper bounds on the order of thickness of the right-angled Coxeter group associated to a graph. Details about his program
are available in his undergraduate thesis, which I supervised,
Algorithm computations of thickness in right-angled Coxeter groups
The following diagram (using data from the program CFS) shows a
sharp transition at density
cn-1/2 in the probability of a random graph being
CFS.
The movie below shows the prevalence of thickness in random
right-angled Coxeter groups;
data for this movie was computed using TEA.