Introduction

At first, I was interested in using networks to visualize the relationships on the Netflix dating show Love is Blind. 30 contestants (15 men and 15 women) on the show go through a period of dating each other for 10 days before possibly becoming engaged, and then potentially get married in 2 weeks. As the contestants form multiple connections with each other on the show, there are often love triangles (or even squares). Sometimes contestants agree to get engaged to one person but later decide to end the relationship and get together with one of their previous “pod” connections.

I was especially interested in visualizing the connections from the most recent season (season 6) since there were multiple love triangles. Also, there was a lot of buzz on social media around one of the contestants who had multiple women vying for his attention. I was interested in seeing the “shape” of Love is Blind connections and potentially comparing the shapes of each season.

Inspiration

I was inspired by “The Chart” on the L word, a hand-written network of dating history between the show’s characters that has been reproduced and visualized repeatedly in different iterations (including from past Info Vis students).

While I am familiar with some projects that work with data from reality dating shows (The Bachelor Data) I have not seen network visualizations of dating shows before.

Creating the Love is Blind network data

I created a spreadsheet of nodes with each contestant either being a source or target in an undirected connection. I decided to assign a weight based on what kind of connection it was – 1 for dated (in the pods), 2 for engaged, and 3 for married. Assigning this kind of categorization to relationships might be problematic, but it makes sense for the structure of the show which is divided into 1. The “pods” dating phase where all contestants date before deciding who to get engaged to, 2. The engaged couples start to live together and plan their wedding, and 3. The remaining couples decide at the altar whether they will get married or break up. Weighing the edges this way also would allow you to see that a contestant may have dated multiple people, but only got engaged (or married) to one person.

I played around with a few network types in Gephi. I set the nodes to be sized based on ranking of degree, so that the contestants with more connections would be sized larger. This allows you to see that Jimmy had the most connections in the pods, and was the center of at least 3 love triangles (2 of which were filmed, and a third love triangle that has since been revealed on social media).

I also experimented with using different colors to show the edge categories, but ultimately decided to move in a different direction with the network. Since a season of Love is Blind will, on average, follow the stories of 5 engaged couples, the amount of connections between contestants is small. I decided to continue exploring networks through a different reality dating show with more connections, Perfect Match.

Perfect Match

Perfect Match is a Netflix reality dating show that casts former contestants from other Netflix reality shows including Love is Blind, Too Hot to Handle, The Circle, The Mole, and more. Contestants are asked to choose another person to match with at the end of each episode and compete in challenges to determine the most “compatible” couple. While the show starts off with 10 contestants, they are given the opportunity to bring in new contestants throughout the show, resulting in a total of 23 contestants.

I created the Perfect Match node data following a process similar to making the Love is Blind spreadsheet; using data from Wikipedia, I added each contestant as a source and who they connected with as a target in an undirected connection. I assigned the weight as the total amount of times those contestants had “matched” with each other. For example, Joey and Kariselle matched with each other all 10 episodes and never matched with any other contestants.

Once I imported the data into Gephi, I experimented with different shapes for the network.

While the Fruchterman Reinhold creates a beautiful shape, I wanted the network to show a clearer story of the data. I found that the Yifan Hu network style created the shape I found most impactful.

I like that this network shape makes it much easier to see the clusters of connections and the outliers. The node sizes are also adjusted to show the degree of connections each contestant made.

I spent a while tweaking the appearance of the network, and started to incorporate the Perfect Match branding colors (the edges here are “Perfect Match” pink). At one point I experimented with showing what original show the contestants came from. I added a column to the data for “Original show” and color coded them, but ultimately I think this made the network more visually messy. Later I realized I also needed to add the contestants Calvin and Colony as nodes; since they appeared on the show but never matched with any other contestants, they would appear as outliers.

I ended up moving the network into Illustrator so that I could incorporate some more design elements. The final draft of the network incorporates the colors from the Perfect Match logo as well as headshots for each Contestant.

Reflections on the Network

Whether you are a viewer of the show or not, the network may reveal some surprising things. Kariselle and Joey only matched with each other throughout the show (and even became engaged towards the end) but they were not voted “most compatible couple” and did not win the show! Dom and Georgia were instead crowned the winners, although they matched with each other 7 times compared to Joey and Kariselle’s 10.

Also, contestants and viewers would often talk about Francesca being the most promiscuous or fickle contestant, but in reality Ines made the most matches with different contestants. Poor Ines tried very hard to find her perfect match, but ultimately left the show after not finding anyone she felt compatible with.

Perfect Match has received some praise for its positive portrayal of bisexuality – in a dating show that is pretty determined to set up male contestants with female contestants, this network shows that it wasn’t completely heterosexual. Francesca and Abbey are both openly bisexual and matched with each other on the show, creating a cross-gender connection between Francesca, Abbey, Bartise, Damian, and Ines. And although it isn’t visualized in the network, Kariselle spoke about her own bisexuality and hooked up with Francesca on the show (although they never officially “matched”).

Future Directions

I would like to continue tweaking the design and appearance of the network. Calvin and Colony’s node sizes should be adjusted for visibility. The Bartise/Izzy/Ines cluster is a little crowded, so I would like to space them out a bit more so that the text doesn’t look so cluttered. I would also like to add some introductory descriptive text and annotations with context, including from the reflections noted above. Since I finished the network in Illustrator, it is not interactive, but I think there are some ways to incorporate some movement, possibly in a gif or a few slides that show one annotation box at a time.