by computing the depth of each node, and adding so...
# data2viz
p
by computing the depth of each node, and adding some more linkDistance (on ForceLink) and/or adjust radius size (on ForceCollision) depending on this depth, I think you can manage the way your nodes place way better.