Is Talking About the “Winter Olympics” Worse for your Social Network than “Valentine’s Day”?

The answer to the above question is clearly, “Who knows?!,” (or maybe “Who cares?!”) but by mining the social networking soup that is Twitter we may be able to approximate a guess—and the answer may surprise you. In keeping with my previous post celebrating America’s Hallmark Holiday, I used the twitter_hash_net.R script available in the ZIA Code Repository to extract the social networks of those tweeting about either “Valentine’s Day” or “Winter Olympics”. With this data, we can compare the cohesive characteristics of these networks in an attempt to answer the question: Is Talking About the “Winter Olympics” worse for your social network than “Valentine’s Day”?

Comparing Network Metrics

We begin by examining some very basic network metrics. It appears from this completely unscientific sample that the Valentine’s Day network is both larger and more cohesive. The key statistics here are the mean degree and mean shortest paths. Though only varying by small fractions, given the difference in overall size (+671 vertices for Valentine’s) we would actually expect the Winter Olympics network to have higher values for these statistics. Also, the largest connected component of the Winter Olympics network has zero transitivity, i.e. is a tree graph, making it extremely sparse. The densities are a bit misleading given that these networks contain several disconnected components.

Visualizing 2-Cores

“Valentine’s Day”	“Winter Olympics”

Next, we visualize the 2-core of the network to reveal the core structure, and this provides further insight into the social cohesiveness of these groups. The Valentine’s Day 2-core contains a single component with significant structural cohesiveness, and two actors (@Kellypetters and @Maiyanicutie) with near perfect structural equivalence. Interestingly, upon examining their tweets, it may be that these accounts are actually the same person. That said, the network on the left is clearly more cohesive than that on the left. The Winter Olympics network are closer approximation of geometric graphs than social ones.

The moral of the story? It may be that contrary to popular theory about the high correlation of loneliness and Valentine’s Day, those looking forward to the Winter Olympics may be a more isolated social group. This does not correctly address the questioned as stated above, as it highly likely that individuals from these groups simply have a different tie generation process.

That said, perhaps some of those Valentine’s Day promoters should get out there and follow the Winter Olympics crowd; or better still, maybe Hallmark should be making greeting cards for the lonely bobsled enthusiasts?

Automatically Generated Related posts:

1. How to: Use Python and Social Network Analysis to Find New Twitter Friends
2. Measuring Network Effects with No Network Analysis
3. Valentine’s Day Special: Top Ten Soc Sci/Math/Stat Pickup Lines EVER
4. Network Analysis in the Social Sciences
5. Learning About Network Theory