Game of Thrones features an ensemble of protagonists, from Jon to Jaime, from Arya to Tyrion. The interwoven story arcs for these myriad protagonists form the unified super-plot that is the Song of Ice and Fire. So rather than asking “Who is the protagonist of Game of Thrones,” we should ask “What is the scale for each protagonist?”
Last year, I collaborated with Michael Chemers, a Professor of Dramatic Literature to write a paper about using data visualizations as a tool for the analysis of vast narratives. I’ve added a page where you can read more about Fractal Protagonists and the strange conundrum of the Kingslayer. Below is an example of a Character Centrality Diagram for Season 7, which offers a simplified view of the centrality metrics, and gives a sense of scale of each character.
This plot shows
PageRank Centrality (important interactions) on the x-axis
Eigenvector Centrality (important connections) on the y-axis
Weighted Degree Centrality (total interactions) in hue (purple is low, yellow is high)
The centralities have been normalized so that they take on values between 0 and 1. By Season 7, all three of these centralities basically agree. This was not true in earlier seasons, when characters were traveling their various story arcs. In fact, the way characters move through this diagram from season to season tells you what kind of narrative journey they are on.
I have posted the analysis for Season Seven. It was filled with convergence: an all-female war council, family reunions for Stark and Lannister (and for Targaryen as well?), and Tormund & the Hound together again for the first time. And the network crowns a brand new #1… what took him so long?
This brings “Network of Thrones” up to date, with every season mapped and analyzed. Check them out using the links to the right.
The holidays gave me some time to come back to this project. I’ve posted the analysis of Season 2. The sophomore season is filled with adversity, especially for the Starks. Winterfell falls to the betrayal of its adopted son. Arya is a hair’s breadth from discovery. Jon is lost to the wildlings. Sansa is helpless and friendless in King’s Landing. Robb and Catelyn find their efforts to rally against the Lannisters hampered by the conflict between Renly and Stannis Baratheon. Meanwhile, Daenerys is desperate and directionless in the Red Waste.
The narrative takes a strong tack towards the Lannisters, and we begin to see the world through their eyes. In particular, Arya’s masquerade puts her next to Tywin Lannister, revealing a formidable opponent. And of course, the misfit Tyrion finds the balance of the realm in his hands, as he checks the ruthlessness of Cersei and the cruelty of Joffrey.
I have added the analysis for Season 6. The network has evolved into three larger communities, two mid-sized communities, and four small satellite communities. Sansa, the Key to the North, may also be the key to the network. Tyrion makes a rebound in Mereen, and Ned Stark reappears with some prominence, thanks to the Three-Eyed-Raven’s flashback tour of the Tower of Joy.
I’ve added my analysis for Season 5, which features Lord Commander Jon Snow, the Sons of the Harpy and an exiled Imp, Jaime’s princess rescue in Dorne, a power struggle between church and state, a girl that has no name in Braavos, and the most unhappy Stark marriage since the Red Wedding. Click through to find out how much things have changed since Bran the Climber scaled Winterfell towers for fun.
The TV data continues to evolve. So have fun with the novels for now while I get that into shape. There are some high quality scripts on genius.com. My thanks to Colby Handy and the fans who have contributed content so far. There are still plenty of scripts missing, if you have a knack for it!
I’ve added the first two networks that we constructed from the TV series (instead of the books). We started with the first season. I admit this information is far removed from tonight’s season premiere. In fact, looking at the complete network for the books might be more satisfying today (even though it is still two seasons behind). We tackled the third season next, for sentimental reasons. That’s the first one we started with all those years ago. Bi
So check out the Season 1 page to whet your appetite, and experience some nostalgia for the simpler times of Ned Stark. Remember when you could tell right from wrong? Yeah, me too. Here’s a preview of the top characters for the first season. No surprises.
Then you can head over to Season 3 to relive some of your favorite memories: the wacky Jaime and Brienne road trip, Ygritte’s confident seduction of Jon Snow, and of course the romance of the Red Wedding. Here are the top performers in the third season. You can’t argue much with the names (okay, except for Joffrey), but you may be surprised by the ordering.
I have added pages for each of the books, as well as a separate analysis that gathers all five volumes into one network. Check out The Novels page to explore the analysis. Here is the combined network for all five books.
And here are the top ranked characters for the combined saga.
I have also added two supporting pages. First, A Primer on Network Science contains the basics that you need in order to ponder the analysis on this site. Second, I have added A Science of Networks, which is an expository article that I wrote for The Skeptic (UK) magazine. This gives a general audience introduction to the fundamental nature of Network Science.
I am actively working on networks for the HBO Series “Game of Thrones.” I will post those networks as I finish them. I should have some up before the new season starts.
Math Horizons is geared towards undergraduate math majors. We hoped to spark some interest about networks in the minds of college students. We were far more successful than we expected, and at a much grander scale. Adam Epstein wrote a charming article for the Quartz about our work. For a week or two, our network captured the attention of Game of Thrones fans worldwide.
The warm reception inspired us to expand on our analysis. I would characterize our original article as a “proof of concept.” We have revisited our work, and brought in two Macalester students (now graduates) to help out. Ari Weiland did some great coding, which greatly improved our process. Rebecca Gold contributed her super-fan insight into the series as we looked to interpret our results. This site contains a summary of our work, and some reflection on the meaning of the numbers hidden in the network.