Example: Slovenian magazines and journals

Over 100.000 people have been asked which magazines and journals they read (survey conducted in 1999 and 2000, source CATI Center Ljubljana). They listed 124 different magazines and journals. The collected data can be represented as 2-mode network:
                 Delo    Dnevnik   Sl.novice ...
Reader1            X                   X     ...
Reader2                     X                ...
Reader3            X                         ...
............     .....   .......   ......... ...

Obtaining 1-mode network

From 2-mode network reader/journal we generated ordinary network, where the vertices are journals
  • undirected edge with value a between journals means the number of readers of both journals.
  • loop on selected journal means the number of all readers that read this journal.
Obtained matrix (A):
                 Delo    Dnevnik   Sl.novice ...
Delo            20714       3219        4214 ...
Dnevnik                    15992        3642 ...
Sl. novice                             31997 ...
........        ......     .....       ..... ...
The second ordinary network on readers would be huge (more than 100.000 vertices) containing large cliques (readers of particular journal).

Data analysis

Because of the huge differences in number of readers of different journals it is very hard to make any conclusions about journals connectivity according to raw data. First we have to normalize network. There exist several types of normalizations:
In all cases each element of the matrix is divided by the proper diagonal element(s).

Geo normalization

Using Geo normalization we divide elements of the matrix by geometric mean of both diagonal elements. In our case this normalization is a measure for connectivity (corelation) among journals.

Geo / with values

MinDir normalization

MinDir normalization makes two things:
  • undirected network converts to directed - the arcs points from the journal with less readers to the journals with higher number of readers
  • the value on the arc corresponds to the percentage of readers of the first journal, who read also second one.
In our case this normalization measures the dependance between journals. The size of vertices corresponds to number of readers. The colors determines weakly connected components.

MinDir / with values