4 ms·

Network visualization: mapping Shakespeare’s tragedies

17 points by hunglee2 11y ago

peter303 11y ago
Someone should graph the Avengers which seems to ave an abundance of main characters. Shakespera results seem to show on or two protangonists and antagonists each works better.
danso 11y ago
Cool analysis...I think the works of Shakespeare are vastly underused for fun little tests of text mining and visualization, given how accessible (well, in the downloadable/publishing sense, if not the language :) ), ubiquitously taught, and relatively standardized they are, as far as works of literature go. But in terms of the article's visualizations, this seems to me a good example of how network visualizations (note that visualization is different from analysis, which are often conflated) are not especially effective, other than to make people want to think "Oooh that's complicated" whether they derived insight or not. In contrast, I find things like this matrix diagram of Les Miserables, as found in the D3 gallery [1], to be much more straightforward, even before you use the dropdown box to interact with it. It's not as attractive or high of "wow" factor, but its information clarity more than makes up for that IMO. OTOH, one thing that the spaghetti network maps do show is that there's much room for more sophisticated analysis. The OP looks at whether characters appeared in a single scene together. I was about to criticize the uselessness of the Hamlet graph: http://www.martingrandjean.ch/wp-content/uploads/2015/12/Shakespeare-Network-Hamlet.png http://www.martingrandjean.ch/wp-content/uploads/2015/12/Sha... But then I realized...that's what it should look like if you are simply doing scene presence analysis. From my memory, Hamlet interacts with a great many more characters than do Gertrude or Claudius, such as the Ghost and the gravediggers. However, technically, the Ghost and the gravediggers appear in all the same scenes that Gertrude and Claudius do. I think the network graph would look much different if it were based on adjacent dialogue (or some other way to distinguish between co-appearance and actual interaction). It's a little more parsing but it would be more accurate in quantifying the strength of the network ties. [1] http://bost.ocks.org/mike/miserables/ http://bost.ocks.org/mike/miserables/

17 points by hunglee2 11y ago

Someone should graph the Avengers which seems to ave an abundance of main characters. Shakespera results seem to show on or two protangonists and antagonists each works better.

danso 11y ago

Cool analysis...I think the works of Shakespeare are vastly underused for fun little tests of text mining and visualization, given how accessible (well, in the downloadable/publishing sense, if not the language :) ), ubiquitously taught, and relatively standardized they are, as far as works of literature go. But in terms of the article's visualizations, this seems to me a good example of how network visualizations (note that visualization is different from analysis, which are often conflated) are not especially effective, other than to make people want to think "Oooh that's complicated" whether they derived insight or not. In contrast, I find things like this matrix diagram of Les Miserables, as found in the D3 gallery [1], to be much more straightforward, even before you use the dropdown box to interact with it. It's not as attractive or high of "wow" factor, but its information clarity more than makes up for that IMO. OTOH, one thing that the spaghetti network maps do show is that there's much room for more sophisticated analysis. The OP looks at whether characters appeared in a single scene together. I was about to criticize the uselessness of the Hamlet graph: http://www.martingrandjean.ch/wp-content/uploads/2015/12/Shakespeare-Network-Hamlet.png http://www.martingrandjean.ch/wp-content/uploads/2015/12/Sha... But then I realized...that's what it should look like if you are simply doing scene presence analysis. From my memory, Hamlet interacts with a great many more characters than do Gertrude or Claudius, such as the Ghost and the gravediggers. However, technically, the Ghost and the gravediggers appear in all the same scenes that Gertrude and Claudius do. I think the network graph would look much different if it were based on adjacent dialogue (or some other way to distinguish between co-appearance and actual interaction). It's a little more parsing but it would be more accurate in quantifying the strength of the network ties. [1] http://bost.ocks.org/mike/miserables/ http://bost.ocks.org/mike/miserables/