Hello!
Facebook has many bi-modal networks within it. User interact with Posts, for example, with Likes and Comments.
Calculating meaningful network metrics for bi-modal networks is difficult. For example, the most central objects in a bi-modal network might be a mix of Users and Posts.
While there are efforts to create metrics that can handle the complexity of a bi-modal network, they all in some way do so by creating a "projection" of a single mode network from the data.
A Network made of Users and Posts can be transformed or projected into a network of just Users connected to Users (by hidden shared connections to Posts) or just Posts connected to Posts (by hidden shared User connections). Once in a unimodal form, these networks can be measured using most of the standard measures of degree, centrality, density, etc.
The NodeXL SocialNetImporter provides for the importation of both the bi-modal User to Post network edges and the transformed User to User and Post to Post networks. Importing just one of the unimodal networks can make it easier to analyze the connections among the Users or Posts on a Facebook fan page.
These metrics will then highlight the key people or content elements that were located in strategic positions in the network. Which people are most connected? Who is most uniquely connected? How do Posts cluster together (because different groups like different kinds of content)?
Regards,
Marc
Facebook has many bi-modal networks within it. User interact with Posts, for example, with Likes and Comments.
Calculating meaningful network metrics for bi-modal networks is difficult. For example, the most central objects in a bi-modal network might be a mix of Users and Posts.
While there are efforts to create metrics that can handle the complexity of a bi-modal network, they all in some way do so by creating a "projection" of a single mode network from the data.
A Network made of Users and Posts can be transformed or projected into a network of just Users connected to Users (by hidden shared connections to Posts) or just Posts connected to Posts (by hidden shared User connections). Once in a unimodal form, these networks can be measured using most of the standard measures of degree, centrality, density, etc.
The NodeXL SocialNetImporter provides for the importation of both the bi-modal User to Post network edges and the transformed User to User and Post to Post networks. Importing just one of the unimodal networks can make it easier to analyze the connections among the Users or Posts on a Facebook fan page.
These metrics will then highlight the key people or content elements that were located in strategic positions in the network. Which people are most connected? Who is most uniquely connected? How do Posts cluster together (because different groups like different kinds of content)?
Regards,
Marc