The One With The Social Networks
Source: https://towardsdatascience.com/the-one-with-all-the-friends-analysis-59dafcec19c5
This article analyzes the scripts of the television sitcom Friends and looks to answer some thought-provoking questions about the show. First, the author seeks to determine the most popular character in the show by examining the number of lines spoken by each character and the number of lines in which each character is mentioned throughout the ten seasons. Rachel gets the top spot with over 9200 lines spoken and Ross comes in a close second at just over 9000 lines. This finding would not surprise a fan of the show, who has seen Rachel and Ross featured as the main plot throughout the series. Ross is also the most mentioned character in the show. Likewise, Monica and Chandler made up a crucial plotline in the show and spoke a nearly identical number of lines, 8403 and 8398 respectively.
Having assessed the role of each character, the author goes on to examine how the characters interact by analyzing the social network among the Friends characters. Specifically, the author looks at how many times a character mentions another character to create a weighted directed graph relating to the characters. Some interesting findings were that Ross mentioned Rachel the most with 622 lines and Rachel referred to Ross the most with 550 lines. Interestingly, although Monica mentioned Chandler the most, Chandler spoke the most of Joey. Instead of constructing the usual weighted directed graph to represent the comprehensive Friends social network, the author utilizes a chord diagram, as pictured below, to represent the strength of the ties between characters in a more visually-appealing manner.
In this diagram, the size of the chord for each character’s section represents how many times they said the name of the connecting character. It is easy to picture a similar undirected version of this network with weighted edges, such as below:
If we view the red and gray edges as strong ties and the purple and black edges as weak ties, we can see that this network satisfies the Strong Triadic Closure Property. Triadic closure is the property among three nodes A, B, and C, such that if a strong tie exists between A – B and A – C, there must exist an edge between B – C. In this case, it means that if one character is particularly good friends with two other characters, those two characters must be friends themselves. This turns out not to be an issue since this is an example of a complete graph where every pair of nodes is connected by a pair of unique edges (all characters in Friends are friends, how appropriate).
Having constructed a social network for the Friends characters, the author looks to determine the most “central” person in the network and does so by computing a centrality score for each character. Using the number of mentions as edge weights, the author calculates eigenvector centrality to measure the importance of each character in the network. This algorithm aims to quantify the influence of a person in a social network, based on connections with other important people. If a node is pointed to by many nodes (which also have high eigenvector centrality), then that node will have high eigenvector centrality. We see that eigencentrality is calculated as follows:
With an emphasis on links with other people, it is easy to see how this algorithm may be applied to larger networks such as Twitter. Using retweets and likes as weights, this algorithm can provide insight into which accounts on the social media platform are the most important. Similarly, a variant of this algorithm is used by Google’s PageRank to rank webpages in their search engine results. PageRank works by counting the number and quality of links to a page and giving each page a relative score of importance and authority under the assumption that more important websites are likely to receive more links from other websites.
The result of applying the algorithm to the Friends social network is shown below:
It makes sense that Joey has the highest centrality score due to his unique standing as the “glue” of the group who always interacts with the other characters. It is also reasonable that Ross and Rachel have a lower centrality score, despite their prominent plotline, since they occupied most of each other’s time during the show. Phoebe, who displayed a distinctly independent personality and had the fewest familial, romantic, or roommate-related ties to the others, has the lowest centrality score in the social network.