- What different types of sociological information are encoded within a web graph?
- Can the web graph help to identify actors within a particular issue space?
- Can the web graph identify the most central actors within an issue space?
- Can the web graph tell us something about the relationship between individual actors within the issue space?
- Can the web graph structure tell us anything about the possible function of individual nodes?
- Can the web graph tell us anything about the importance of a given issue relative to other issuess that are discussed and deliberated on the Internet?
- Can the web graph tell us anything about the broader orientation of individual issues within larger virtual social communities?
Web graph research is currently at the exploratory stage. The bulk of our work involves generating graphs of targeted issues (for example: progressive left politics, electronic voting, privacy) and interpreting them. The majority of these graphs are static graphs, showing a snapshot of the network at a single point in time. We are also beginning to explore dynamic mapping, looking for signs of network growth and change over time.
In addition to this general exploratory work, we have been collecting data to answer questions about the relative objectivity of web graphs. Specifically, how dependent is the web graph on the seed URLs that are used to generate it? Some critics of web graph analysis have argued that observed network attributes, such as centrality, are determined by the chosen seed URLs and thus have little research value. Our research so far suggests that this criticism is unwarranted.