0
Simplifying Social Network Visualizations<br />Lynn Cherny, Ph.D.<br />
Why Care? Network (data) is everywhere now…<br />1.<br />Lynn Cherny, Ghostweather R&D LLC<br />
Lynn Cherny, Ghostweather R&D LLC<br />A social network is a social structuremade up of individuals (or organizations) cal...
Lynn Cherny, Ghostweather R&D LLC<br />And then Moritz Stefaner posted this during the summer…<br />
How can we untangle the hairball?<br />Using social network analysis metrics and filtering the data set…<br />Lynn Cherny,...
Tools I Used<br />Lynn Cherny, Ghostweather R&D LLC<br />(free) R<br />(free) Python  (NetworkX)<br />(free) Gephi (www.ge...
Lynn Cherny, Ghostweather R&D LLC<br />Gephi – Network vis and analysis<br />Statistics Tab<br />Color -By…<br />Graph Pre...
Calculate Stats (Gephi and NetworkX)<br />Lynn Cherny, Ghostweather R&D LLC<br />
Some Important Stats – Degree, Authority, Betweenness…<br />Lynn Cherny, Ghostweather R&D LLC<br />
R – find correlations among stats<br />Lynn Cherny, Ghostweather R&D LLC<br />
Lynn Cherny, Ghostweather R&D LLC<br />
Lynn Cherny, Ghostweather R&D LLC<br />
Betweenness<br />Lynn Cherny, Ghostweather R&D LLC<br />A measure of connectedness between (sub)components of the graph<br...
Lynn Cherny, Ghostweather R&D LLC<br />
Use “Degree” to scale size of node label in Gephi… still a mess!  Not enough yet.<br />Lynn Cherny, Ghostweather R&D LLC<b...
Detecting Sub-Communities-and then filtering.<br />Lynn Cherny, Ghostweather R&D LLC<br />
Community Detection Algorithms<br />Lynn Cherny, Ghostweather R&D LLC<br />Louvain method, implemented in both Gephi and N...
Lynn Cherny, Ghostweather R&D LLC<br />
Lynn Cherny, Ghostweather R&D LLC<br />
Lynn Cherny, Ghostweather R&D LLC<br />“The Authorities”<br />“The Researchers”<br />
Summary<br />Lynn Cherny, Ghostweather R&D LLC<br />You can go from a hairball to an analysis…<br />With free tools (plus ...
Lynn Cherny, Ph.D.lynn@ghostweather.com<br />Lynn Cherny, Ghostweather R&D LLC<br />Need data help?  Have a hairball?  Con...
Upcoming SlideShare
Loading in...5
×

Simplifying Social Network Diagrams

3,510

Published on

Some tips on how to simplify a hairball network diagram using social network stats, especially in Gephi (or NetworkX).

Published in: Technology, Education
0 Comments
7 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total Views
3,510
On Slideshare
0
From Embeds
0
Number of Embeds
10
Actions
Shares
0
Downloads
85
Comments
0
Likes
7
Embeds 0
No embeds

No notes for slide

Transcript of "Simplifying Social Network Diagrams "

  1. 1. Simplifying Social Network Visualizations<br />Lynn Cherny, Ph.D.<br />
  2. 2. Why Care? Network (data) is everywhere now…<br />1.<br />Lynn Cherny, Ghostweather R&D LLC<br />
  3. 3. Lynn Cherny, Ghostweather R&D LLC<br />A social network is a social structuremade up of individuals (or organizations) called "nodes", which are tied (connected) by one or more specific types of interdependency, such as friendship, kinship, common interest, financial exchange, dislike, sexual relationships, or relationships of beliefs, knowledge or prestige. [--Wikipedia]<br />Postdoc at Indiana University ad:<br />“Areas of focus will include information diffusion patterns, epidemic models for the spread of ideas, interactions between network traffic and structure dynamics, and agent-based models to explain the emergence of viral bursts of attention. <br />
  4. 4. Lynn Cherny, Ghostweather R&D LLC<br />And then Moritz Stefaner posted this during the summer…<br />
  5. 5. How can we untangle the hairball?<br />Using social network analysis metrics and filtering the data set…<br />Lynn Cherny, Ghostweather R&D LLC<br />
  6. 6. Tools I Used<br />Lynn Cherny, Ghostweather R&D LLC<br />(free) R<br />(free) Python (NetworkX)<br />(free) Gephi (www.gephi.org)<br />Tableau Desktop<br />Excel<br />Illustrator and Photoshop<br />
  7. 7. Lynn Cherny, Ghostweather R&D LLC<br />Gephi – Network vis and analysis<br />Statistics Tab<br />Color -By…<br />Graph Preview<br />Layout Method<br />Filter Rules<br />
  8. 8. Calculate Stats (Gephi and NetworkX)<br />Lynn Cherny, Ghostweather R&D LLC<br />
  9. 9. Some Important Stats – Degree, Authority, Betweenness…<br />Lynn Cherny, Ghostweather R&D LLC<br />
  10. 10. R – find correlations among stats<br />Lynn Cherny, Ghostweather R&D LLC<br />
  11. 11. Lynn Cherny, Ghostweather R&D LLC<br />
  12. 12. Lynn Cherny, Ghostweather R&D LLC<br />
  13. 13. Betweenness<br />Lynn Cherny, Ghostweather R&D LLC<br />A measure of connectedness between (sub)components of the graph<br />http://en.wikipedia.org/wiki/Centrality#Betweenness_centrality<br />
  14. 14. Lynn Cherny, Ghostweather R&D LLC<br />
  15. 15. Use “Degree” to scale size of node label in Gephi… still a mess! Not enough yet.<br />Lynn Cherny, Ghostweather R&D LLC<br />
  16. 16. Detecting Sub-Communities-and then filtering.<br />Lynn Cherny, Ghostweather R&D LLC<br />
  17. 17. Community Detection Algorithms<br />Lynn Cherny, Ghostweather R&D LLC<br />Louvain method, implemented in both Gephi and NetworkX<br />Outputs a classification for each node in the list – 1, 2, 3… corresponding to the “community” they are closest to.<br />http://en.wikipedia.org/wiki/File:Network_Community_Structure.png<br />
  18. 18. Lynn Cherny, Ghostweather R&D LLC<br />
  19. 19. Lynn Cherny, Ghostweather R&D LLC<br />
  20. 20. Lynn Cherny, Ghostweather R&D LLC<br />“The Authorities”<br />“The Researchers”<br />
  21. 21. Summary<br />Lynn Cherny, Ghostweather R&D LLC<br />You can go from a hairball to an analysis…<br />With free tools (plus some standards or extras to make the results pretty)<br />Network analysis is another form of data literacy now – think of these methods as similar to other stats, but specific to graphs.<br />
  22. 22. Lynn Cherny, Ph.D.lynn@ghostweather.com<br />Lynn Cherny, Ghostweather R&D LLC<br />Need data help? Have a hairball? Contact me.<br />See my blog post on this for more info and links:<br />http://blogger.ghostweather.com/<br />
  1. A particular slide catching your eye?

    Clipping is a handy way to collect important slides you want to go back to later.

×