MCSN Tuesday, 20-Sep-11

From CCE wiki archived
Revision as of 10:35, 20 September 2011 by Michaelf (talk | contribs)
(diff) ?Older revision | view current revision (diff) | Newer revision? (diff)
Jump to: navigation, search

Course mechanics

  • Please bring in your laptops so we can work together.
  • Quiz on Thursday, first 10 minutes. Will test your knowledge of concepts from chapters 1 and 2 only through today's reading. Will not be difficult and will not involve Pajek at all.
  • Readings: by section, not page (section 2.4 means read all of it, i.e. 2.4.1, 2.4.2, etc.)
  • Moodle
    • issues?
    • grading for 9/13 done - please see me if you have questions. A couple of people perhaps had trouble using the Moodle...
    • please upload all assignments to the Moodle, including examples prepared for class: "come to class prepared with a Pajek example drawn from the Web, and illustrating chapter content." That way I have everything on file.

Last week...

  • Further brainstorms?
  • Examples of network graphs from last Thursday
  • Assignment 1.7. Examples. What sorts of methodological or ethical issues arose? How do you wish you'd changed your question?

Pajek and SNA through 2.4

  • Important to get everything working. What didn't work?
  • Chapter 2 overview:
    • World system (Immanuel Wallerstein): core, semiperiphery, periphery. Claim: relation between trade position and mode of production. Network data helps establish trade position using block modeling (3 positions).
    • Data: Contains inter-country trade in metal manufactures, in $1k USD units as a valued digraph. Also assigns world system positions from 1980 and c. 1994 using partitions, as well as providing a continent partition, and GDP data as a vector.
    • Graph + additional info (properties of lines and vertices) = network.
    • Often in SNA we want to relate graph properties to properties of lines and vertices. For instance, what is the relation between gender (an vertex attribute) and degree (a network attribute)? Pajek codes additional information as line data (values) and vertex data, the latter of two types: (a) integer classes, and (b) real values.
  • Pajek files
    • Network: .net
    • Partition: .clu file. Consists of multiple classes, coloring the vertices.
    • Vector: .vec. Assigns each vertex a real value, which can be displayed as vertex size.
    • Pajek file: combines the above in a single file
  • Using partitions
    • Examine the three "world trade" partitions
    • A partition assigns every vertex to one and only one class. Usually classes contain more than one vertex.
    • Info-->Partition provides statistics.
    • Marking partitions: with color or labels
    • Moving a class: drag near vertex.
    • Create your own: Partition-> Create null partition, then edit
    • New views, using partitions
      • Local view: extract South America (class 6),using Operations menu and continent partition. Seeing world system positions: Extract second from first.
      • Global view: Operations-> Shrink network. Remove lines with value under $30,000k. Note new labeling of vertices - change them by editing the partition.
      • Contextual view: Operations-> Shrink network (except class 6). Remove lines with value under $10,000k
  • Individual help while others generate examples using techniques in chapter 2.