visualizing a network over time using heatmaps
Most social networks change over time. New social actors enter the network while others leave. Or new network ties are formed, while others are maintained or dissolved. These changes in social networks are at the heart of a vivid stream of social research. To better understand how social networks change (i.e. why are network ties dissolved? What happens when actors leave the network?) researchers have used a host of formal techniques. However, sometimes it is very helpful to start by visualizing social networks and try to observe what is going on. In this post I’ll show how to plot changes in network ties over time for a fixed set of nodes (social actors) using heatmaps. I have created a network_data sample which contains 15 unique nodes observed over 37 time spells (months). In each spell the ties between the two nodes in a dyad emerge, are maintained, or dissolve. I have used R to plot a network for each spell, save it as a pdf file and paste the files together so that one can go through each of the pages by using the cursor on one’s computer. Here is the code. The resulting file should look like this. Basically, each tile in the heatmap represents the tie intensity between a dyad pair. This representation works well with small datasets with stable sets of nodes (stability may be artificially imposed by setting the tie weight to zero for actors leaving the network or actors not yet in the network).