Information visualization researchers developed over the years techniques to help better understand unstructured and structured data present in several application domains. The visualization of these data structures is challenging, also because these may typically evolve over time. This is also particularly the case with network data structures, and dynamic networks in general. Existing techniques may provide insights and facilitate the analyze of networks from multiple points of views. One issue is that the user has to strictly choose a visualization technique over another one that may be be better for some tasks. In practice, each visualization technique have their own potential strengths and limitations.

To facilitate the exploration of such data, it could be useful to use multiple approaches at once. One recent strategy towards the development of optimal visualizations is the combination of techniques (also sometimes called hybrid or composite visualizations). Few of these techniques may also be used in the context of network exploration, and so more work is needed.

Although may factors may affect visualizations, it is generally accepted that the exploration of networks typically require the usage of a variety of techniques, layout algorithms and is completely data and task dependent. Furthermore, interaction techniques such as highlighting can also help reduce the mental effort required to perform tasks, and improve error-rates.

So, an open question is : How does one pick the proper visualization in a given context ? Typically, the practical usefulness of approaches may be verified over extensive empirical studies and in presence of real users. However, this may also constraint the technique to some very specific use cases, making less difficult to use in other contexts, and the needs of users may also evolve.

Designers does not have precise guidelines to guide the design of dynamic network visualizations techniques regarding, in particular, the visual composition strategy, the user type (expert, non-expert, trained user). Empirical studies with involving kinds of users should be pursued to evaluate the applicability in specific fields and contexts.