Understanding Large Temporal Networks and Spatial Networks: Exploration, Pattern Searching, Visualization and Network Evolution

AppendixData Documentation

Here, we document our data sources, our data processing steps, and the construction of the network datasets analyzed in this book. This documentation is organized by the chapter sequence where these datasets are analyzed. The longest description is for the football data because they were constructed from scratch from many sources.

A.1 Bibliographic Networks

We start our description with some general comments on data sources for bibliographic networks before dealing specifically with the scientific networks considered in Chapter 4.

From special bibliographies (BibTE X) and bibliographic services including Web of Science/Knowledge, Scopus, CiteSeer, Zentralblatt Math, Google Scholar, DBLP, Math Sci and others, it is possible to obtain data about works (papers, books, reports, etc.) on selected topics. A typical description of a work contains the following data: authors; title; publisher/journal; publication year; and pages. In some sources, there is additional data including languages, classification of documents, keywords, authors' institution/country affiliation, lists of citations, and abstracts. These data can be transformed into a collection of compatible two-mode networks on selected topics: works × authors; works × keywords; works × countries; and other pairs of characteristics describing works. Besides these networks, we can also get partitions of works by their publication years, partitions of works by journals, vector of number of pages, and, ...