Pull meaning from social networks without reading every post

Recognizing that if you're reading our blog or our social media posts, it’s unlikely that you're attending Dr. Nolker’s presentation at the Sentiment Analysis Symposium, we’ve posted a paper on our site that you can download for free entitled, “Social Computing and Weighting to Identify Member Roles in Online Communities.” This paper was the genesis of what has become a groundbreaking approach for pulling meaning out of social networks. 

For those that would rather get the meat without sifting through the paper, here’s a summary of the paper and how it's useful when doing social network analysis: 

A. Not everything that everyone says in an online social network is worth analyzing. We've all met that guy or gal that post things that don't matter. 

B. Structure mining provides a means for finding and weighting which individuals are most worth analyzing and which individuals we should ignore in an automated fashion.  

C. Not surprising to those that work in teams, the most important people to analyze are Influencers and Motivators (defined more specifically in the paper).  

D. You can detect roles (like Influencers and Motivators)

in online communities (like Facebook, LinkedIN, Twitter, or other more specialized forums like those for Hackers) and sift out individuals that detract from a community by measuring things like:

  • The number of one and two way conversations, 
  • Whether those conversations or posts are directed at individual persons, 
  • The number of different people users converse with, and 
  • How close (first, second, or third level connections) a user has in a given social network, among others.

C. This type of analysis can help businesses target to whom they market, social networks measure how healthy their communities are, and data scientists choose whom to target for more in-depth sentiment, natural language, or link analysis. 

Download the paper for more detail.