Deriving an Interest Graph for Social Data

It is highly important to understand one’s audience for effective communication, especially on the internet where brands spend large amounts of time cultivating their image. Traditionally done with demographic data, the new way to discover who is consuming content is through understanding clicked on topics and predilections. If we take into account url reader’s attitudes and expectations, we can create grouping around opinions. These groupings relate users along meta data to create personas that effectively describe the use-cases for brands. Using these user defined topic centric groups, we can understand the web as a graph and gain valuable insight into community discovery and sociological phenomena which characterize group evolution.

In this talk I will describe how to create meaningful clusters of users of social media. Using bitly data which spans all social networks and more, we have a unique view into what people click on in the internet. We reveal topics and keywords that gain the most attention. Looking along lines of topics, keywords and domains from urls consumed by users that click on bitly links, we can discover the composition of domains and users focus to create distinct and meaningful groups for social communities.