Two and a half hours ago, I was struggling through the Gephi parameters trying to understand the logic of the csv files. After playing around with a few ideas, I decided that I'd found the perfect candidate for social network visualization: Venmo.

Venmo is a super interesting social media network because 1. It's not an ego network and 2. It translates monetary transactions (an easy selection for "edges") into social transactions via a newsfeed. I decided to go through just the first 30 or so venmo transactions within my network of friends on the app, assigned edge labels to types of transactions (food, gas, bills, other), and dropped the spreadsheets into ForceAtlas2 in the hopes that it would show me some interesting friend clusters. I also threw in a gender value for good measure. As expected, the results showed me who the major "friend wallets" in my network were, as well as the most popular types of transactions used via venmo. Certain friend clusters tended to use venmo for different reasons—some showed preference for more straightforward transactions (food, rent, etc) while a network of my comedian friends preferred more creative labels (hence the "other"). 

The ForceAtlas visualization was good with looking at clusters, but perhaps less clear in terms of transactions, so I tried again with a different model. Added in some cosmetic changes, and...voila:

 My conclusion? The Gephi+Venmo combo is incredibly useful, albeit incredibly creepy. But then again, maybe the creepiness factor comes from the novelty of Venmo itself, rather than the visualization. I found myself considering the obscure way that people labeled their own transactions (the "other" category) as proof that this sort of network is organically a one-to-one form of communication. It is implied that people look at the newsfeed, but somehow there's a desire to conceal the fact that this information is broadcast to a public. The result is single word communication and a widespread but unacknowledged practice of ignoring of the "Like" button underneath each transaction.