Talk:Social Networks

From Math Images
Jump to: navigation, search

--Chris 12:09, 14 August 2013 (EDT)Here are my latest edits. I made all the changes I could that didn't need your input.

  • Basic Description: P1S1: The word "innocent" is typically not used to refer to something as sophisticated as this image. Rephrase.
    • Degree Centrality: You introduce "index of exposure" but don't explain it.
  • More Mathematical Explanation

The chart should go before the adjacency matrix. It's not clear what the matrix means without it.

    • Eigenvector C..: The words "overrightarrow x" and other "overrightarrows" appear in the text.
    • Closeness and Bet…

Chris's score is to three digits, all others are to two.

    • Betweenness: I made some changes to the paragraph beginning with: "As you can see…" The most confusing part is the one about i, j, k not being equal to each other. Could this paragraph simply be the last two sentences?
    • Bernoulli
      • S1: suddenly using the word "nodes" instead of "vertices." Clarify this.
      • After equation, the word "proportion" is used incorrectly
      • P6,S1: What does "and otherwise random" mean?
    • Image not helpful for Configuration Model: Figures of different degrees look the same.
    • Basic Reproduction Number
      • Mean Degree <k> LaTex problem?
      • After definitions, you use basic reproductive number. Is it basic reproduction number or basic reproductive number?
  • Why is it interesting? This needs work for the future.

General Feedback

Messages to the Future

  • Adding a discussion of the analogy of the VP-CEO structure, in the basic description, highlighting that CEO's have high eigenvector centrality.
  • More images to the basic description.
  • Add more ideas to the "why it's interesting", there are plenty!
  • Add proofs to the basic reproduction number and Perron-Fronbenious theorem.

References and footnotes

  • Haven't cited much work

Context (aka Generating interest aka Who cares?)

  • Lot's of context is given in the basic description as a motivation as to why study social networks.
  • Why it's interesting section motivates the reader to think about SNA and epidemics and Google PageRank.

Quality of prose and page structuring

  • There are subsections of sections and opening sentences introducing the topic.
  • The sections are motivated, like transitioning to measurements to predictive models.
  • Heavy math is mostly omitted.

Integration of Images and Text

  • Images are referred in the text.
  • Text points to look at what part of the pictures to look at, like with the closeness, betweenness centrality as it outlines the different paths.

Connections to Other Mathematical Topics

  • Links to eigenvectors, matrices and other pages of linear algebra.

Examples, Calculations, Applications, Proofs

  • Examples of applications of formulas in all centrality measurements in MMA.
  • There aren't many proofs.

Mathematical Accuracy and Precision of Language

  • Statements, equations, and mathematical terminology are free of errors.
  • Statements are made as precise as they can be without overwhelming the reader with too many words or dense symbols (and are always precise enough to rule out major ambiguity).
  • Any mathematical term that the reader can't be expected to know is defined in the body text or via a mouse-over or link to another web resource or a helper page.


  • Text is presented in short paragraphs separated by images.
  • No awkward chunks of white space.
  • No image in one section of a page vertically aligns with the text or heading of a different section.
  • The page has been viewed at a few different window sizes to make sure funky things don't happen.

The last feedback I gave to Jason is highly detailed and so would clutter up this page. You can view the document containing the full feedback here.

-Diana 13:17, 15 July 2013 (EDT)

Social network language; connecting measures and graphs: One comment I didn't make last night: it might be useful to make it clear that social networks have been around from early society, while acknowledging that modern digital networks have changed their visibility and dynamics.

I made a comment last night that it is useful to help the reader connect the measures to the features of the graphs. Similarly, it might be useful to use some more of the language of social network analysts such as "cliques" and "entrepreneurial nodes".

-Steve 13:10, 11 July 2013 (EDT)

Basic Description

Regarding the definition of social networks that you give in the fourth paragraph, "agents" is a bit confusing. It might be a technical term I don’t know, but it’s strange here with no prior use of it on the page. Is there any reason not to say “people,” since we’re talking specifically about social networks?

I think you can move all the information from end of the intro, starting with, "Of the topics explained above..." down into the "Centrality Measures" section and work it into the first few paragraphs there. It seems redundant as it is now.

-Diana 12:11, 17 July 2013 (EDT)

Centrality Measures

Degree Centrality

This might seem minor, but you should use something like "probably" rather than "hopefully" in "The characters in the image to the right will hopefully seem familiar." People from other backgrounds really just might never have come across Hamlet.

-Diana 12:11, 17 July 2013 (EDT)

Eigenvector Centrality

In the first paragraph, "it means that it has lots of connections" isn't totally clear. (There are too many "it"s.) Maybe something like, "means that it leads to a lot of other connections"?

-Diana 12:11, 17 July 2013 (EDT)

Closeness Centrality

Betweenness Centrality

A More Mathematical Explanation

I like the table format that you use for the adjacency matrix, in that it makes the connection to real people very clear. However, you should also include a more standard notation of the matrix within brackets and without the labels or cell lines.

-Diana 12:11, 17 July 2013 (EDT)

Hey Jason, I fixed the math syntax error, trying to change your original as little as possible, but I'd suggest writing it this way:

a_{ij}=\left\{ \begin{array}{rcl} 1 & \mbox{for} & i, j \mbox{ adjacent}\\ 0 & \mbox{for} & i, j \mbox{ non-adjacent} \end{array}\right.

-Diana 16:50, 10 July 2013 (EDT)

I altered your code slightly to get rid of the glitching that was happening around the top of the page and the main image.

-Diana 12:35, 11 July 2013 (EDT)

Degree Centrality

Eigenvector Centrality

Closeness and Betweenness Centrality

Dijkstra's Algorithm

Configuration Models

Basic Reproduction Number

Why It's Interesting

Food Webs

Facebook and Google