Longitudinal & Dynamic Networks future

[mbastian]

This post is eager to start the discussion about future features about longitudinal and dynamic networks in Gephi.

What we have in 0.7:
- Each node or edge can have a start and end date as an attribute. These time intervals can be used in filters to use the Timeline component. Users set a time range and the network shown on screen displays nodes and edges inner this period.

What is coming in 0.8:
- Dynamic attribute values. Several values can be set for an attribute, with different time intervals.
- Dynamic API and metrics. It's a Google Summer of Code project currently in development, see specifications here. The aim is to provide a complete API for querying dynamic networks and see metrics evolution over time (degree, centrality, ...).

What we identified as critical:
- Dynamic Layout, what is the best layout to see network evolution?
- Animation, play and export movie
- Slice import. Build longitudinal network from several files
- Your ideas about diffusion ?

We need more references and advices about dynamic layout and longitudinal support in general. Specifying and planning is already half of the work and we will implement (or help you to do it) what users need the most first.

[ciro]

Hi! I think it would be important to allow edge-oriented longitudinal analysis. In many interesting longitudinal graph sources, one is interested in the temporal properties of edges. For example, if edge 1 <-> 2 goes on and off a certain number of times during the evolution of the graph, it would be important to have a straightforward way of computing that quantity (to map it to edge color in a layout, for example). I don't know how possible/easy that is with the current functionality for dynamic graphs.

[admin]

An interesting approach from the Vis Stanford Group on using interactive Timeline: http://hci.stanford.edu/research/isis/

Events are visualized as timelines which use screen space efficiently. A user explores the data set by interacting with existing timelines to create new timelines with related events. This provides the user with an exploration history, which allows users to backtrack and to explore multiple paths.

[panisson]

I see the study of proximity networks as one major application for Dynamic networks in Gephi. Proximity data can be seen as a dynamic network of contacts between devices, with different contact durations. Contact time and inter-contact time distributions are some of the metrics subject to study in this type of networks. Data can be generated by real world experiments (such as http://www.sociopatterns.org or http://reality.media.mit.edu) or by different mobility models. The study of different protocols for information diffusion in these networks is also an interesting subject (http://en.wikipedia.org/wiki/Delay-tolerant_networking).
It's difficult to choose a good layout to see the evolution of these networks, as the nodes are disconnected in most of the time, and there is no main connected component.