Networks and trees

Networks and trees

1. Connection (link marks)

   • Idiom: Force directed placement
     • Placement is a consequence and is often non-deterministic
     • Repositioning could be distracting if there is multiple views
     • Too many nodes and edges can create occlusion
   • Link density
   • Hops

2. Matrix views

   • Idiom: adjacency matrix
     • Only in directed graphs the whole matrix has a meaning. Undirected graphs will form a triangular matrix
     • Are highly effective for very high information densities
     • Predictability, stability (adding more information causes small visual changes) and support for reordering
     • Unfamiliarity, user training is needed

3. Cliques

4. Containment: hierarchy marks

   • Idiom: Treemaps
     • Perform better for understanding attribute values at the leaves of the tree
     • Poor for tasks focused on the topology of the treees