That Is Not A Bookmark

Loranne's Adventures in Librarying

GoT Family Tree… of sorts

For this semester’s Information Visualization class, we were assigned the following:

“Create a visualization that depicts the various families, factions, and allegiances in the Game of Thrones series. This visualization should go beyond being a family tree to include visual encoding(s) that will allow us to see alignments and other types of associations.”

I wanted to try to do something more computational with rather subjective data, and I kind of lucked out in that my first idea worked (for once). I created a spreadsheet, with my top 50 most memorable characters (based on only the knowledge I have from watching the two seasons of the HBO series). I later cut it down to 47, due to three of those characters (Robin Arryn, for instance) being children or less prominent

I placed the character names on both axes, and created my scale, of -5 to 5. For example, Bran Stark and Theon Greyjoy hate each other -5 ( I refuse to offer spoilers as to why), whereas Cersei and Jaime Lannister love each other +5 (do they ever!). It was interesting to have to think of such dynamic relationships this way: reduced to a number.

The end result is something like a heatmap. By creating a spreadsheet this way, I hoped to see patterns develop, and I did. There are interesting clusters where people who don’t even really interact ever (thus far) are all centered around a common friend (or enemy). Sadly, I wasn’t able to execute the interactive chart I’d initially envisioned. I’d like to be able to edit the sort criteria at any time. However, I can still do this in Excel, which is kind of fun.

Classmate and rockstar Ben Chartoff has created an interactive version of my heatmap using Processing. It’s pretty great (and neater, too).

I reviewed my data, and, rather than mirroring relationships (for example, Sansa’s row intersecting with The Hound’s column was previously set to the same value as The Hound’s row intersecting with Sansa’s column), rows now represent how the name on the row feels about the name in the column. This allows for more accurate representation of unequal relationships, where before I used an average for uneven affections.

3 comments on “GoT Family Tree… of sorts

  1. bsegovia
    July 11, 2013

    Very awesome! I Got here through a Reddit post to Ben’s interactive map. this is definitely one of the coolest GoT projects I’ve come across! Kudos!

    I’m really curious about your data collection process for this project. Did you just use the show’s first two seasons as reference material for the “top 50” most memorable characters while using the books as your basis for the relationship dynamic? Was your +5/-5 rating based on your subjective interpretation of that relationship or an average score from many responses? …or some other method?

    I really love this!


    • Loranne
      July 12, 2013

      Thank you! I’m glad you enjoyed it! It was really fun to make, and seeing the end result of Ben’s interactive version just made my year.

      To answer your question – I used very subjective data for this, which I completely generated on my own. When I made this, I hadn’t yet read the books (working on that now!), so I used my own personal understanding of the relationships between these characters, at a specific point in time – the end of the second season of the show.

      In my original draft, which is the one featured in this post, I averaged certain relationships. For instance, if you know the story, you know that Sansa Stark and Sandor Clegane feel very differently toward one another. At the end of Season 2, Sansa is still scared of him, but he definitely likes her. So the number I used for them was an average of what should have been two very different values. This also means that this original image is a perfect mirror along the self-to-self relationship axis (the black line).

      Once Ben and I determined that he was going to rework it into the awesome interactive version, I revised my data, and, going row by row, adjusted values to reflect how the character whose name labeled that row felt about the character in the intersecting column. That helped me to resolve this and other similarly… uneven affections (or hatreds).

  2. Mahesh CR
    July 12, 2013

    Loranne, brilliant work..both yours and Ben’s. Big fan of GoT. Have a question for you. How did you extract character relationships? I see that you say “top 50 most memorable characters” but was not sure if you did some text processing to extract characters and nature of their relationships.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


This entry was posted on March 31, 2013 by in portfolio.
%d bloggers like this: