“How you doin'?” —Joey Tribbiani, FRIENDS character



Network Analysis of Social Connections

Using a dataset of character lines from F.R.I.EN.D.S and The Office, can we uncover hidden interaction patterns in our favorite sitcoms? In the Network Analysis of Social Connections data sprint, JC guided Eskwelabs Data Club members in using the techniques of network science to create interactive visualizations that show complex social narrative structures, which may shed light on how these sitcoms became so popular.

JC Peralta, Network Analysis of Social Connections Sprint Instructor and Data Scientist at AC Energy Corporation (an energy platform of the Ayala Group). || Interviewed by Francine Tan, Eskwelabs Associate Communities Manager

Why did you propose the topic of network analysis of social connections? What's the rationale or story behind this sprint?

I really like doing data projects out of things we do out of leisure. So topics like music, tv shows, movies are always on top of my head whenever I think of something to teach. Data science and programming can be quite intimidating for beginner learners, and I feel like talking about things we already like makes it more fun and interesting.

For this sprint, we talked about two of the most popular and successful sitcoms of this generation: The Office and F.R.I.E.N.D.S. Unlike thriller and drama series, what makes these sitcoms so addictive lies not so much in their plot but mainly in their characters and how they interacted with each other. These interactions might seem spontaneous when we're watching them, but if you take a step back and take a look at the bigger picture, each of these interactions contributed to the relationships formed by the characters that, in turn, formed the entire show's dynamics and brand of humor that we enjoyed and loved. This setup makes sitcoms a perfect subject in learning network theory using Python.

What did the Data Club members do while participating in your data sprint?

Simply put, network theory is a study of relationships of any given set of objects. These networks are represented as a graph where objects are represented by dots called “nodes” and a line called an “edge” is drawn between any two related dots. With only these two simple fundamental rules, network theory has already produced a wealth of insights in genomics, chemistry, transportation, and even sociology and politics.

Using a dataset of character lines in the sitcoms, we used the techniques of network science to create an interactive visualization that show the complex social narrative structures in the show. The characters are the nodes, and an edge is drawn in between any two of them if they have been together in a sufficient number of scenes. The size of the nodes tells how well-connected a character is, and the colors show how the network algorithm clustered them into groups.


To be able to do this, we had spent sessions on wrangling and basic visualizations, the basics of network theory, a method to determine character clusters, and some network metrics to gauge how important a character is at a particular season in the sitcom.

We actually went beyond what has happened in the sitcom and explored hypothetical situations where we tried to determine the resulting effect of doubling the interactions of a particular character, and even removing a character entirely.

How did you make this data sprint fun or engaging?

We had a "Who said it?" Kahoot quiz to see how well they remember the lines in the sitcom (and also to know who the top fan is among the class!).

The class formed groups according to their sitcom of choice and at the end of the sprint, they shared their findings to the whole class by delivering a short presentation.

What are the applications for this in the future?

It may seem like this sprint is tailor made for just sitcom data analysis but the network theory methods they learned here are applicable for any use case that you can make a network out of. To share a few examples, these skills are needed to build recommender engines in ecommerce, detect communities and influential accounts in social media, tracing the effect of new drugs at the molecular level, and so on. The possibilities are endless!


Want to participate in JC’s Network Analysis of Social Connections data sprint for FREE from April 26-May 7?Join Eskwelabs’ Data Club and become a member today to be able to participate.

No worries, you can join with a free trial first if you’d like!


About Data Club

A virtual upskilling experience as a hands-on laboratory where you are guided by industry mentors to build data projects with friends. Lifelong learners at different levels of data proficiency are welcome!

Learn more about Data Club here.

The future of work is data. Eskwelabs creates opportunities for people and teams to thrive in the future of work through data skills education.

Updates from Eskwelabs

Subscribe to our newsletter receive updates and insights on future of learning.
For admission enquiries, please contact admissions@eskwelabs.com.

All Rights Reserved. Copyright © Eskwelabs 2021.