Introduction

The National Basketball Association (NBA) is the premiere and most watched professional basketball league in the United States. Over the past two decades, there have been dozens of superstars in the NBA whose superior athleticism, will to win, and ability to garner support of fans worldwide is unprecedented. This network visualization demonstrates both the biggest stars of this period, the relationships between NBA all stars and the players who they have teamed up with in an all star game, and the shifts over this era among who are the biggest stars in the league.

As a fan of the NBA who was born in 1995, these players have played a pivotal role in shaping this league drastically over the years of which I have followed it closely. I have watched in awe as Bryant, Garnett, Shaq and Wade dominated; as Duncan, Pierce, Paul, and Gasol led; as LeBron single-handedly changed the way in which NBA stars, and athletes at large, negotiate their contracts and their agency as shapers of culture. 

Materials

I was inspired by the Les Misérables Character Network Visualization by previous Pratt Institute visualization student nfesta. This visualization impressed me in that it very beautifully demonstrates a complex web of interaction between a large number of characters, and the classification between groups allows for easy dissemination of who most regularly interacted with whom.

I gathered my data from Data World, which was uploaded two years ago by Gabe Salzer. The data lists all NBA all stars from 2000 to 2016, with additional information such as the team they played on that season, their position, their height and weight, as well as when they were drafted and what country they are from.

Methods

Using code language created by Professor Chris Alen Sula, I brought the dataset into R as a CSV file and converted each player to a string; based off which and how many all-star teams they played on. These strings became the basis of the nodes of this visualization, each all star player in this era and the number of teams they played on; as well as the edges, which are the teammates they played with.

As you can see above, Gephi populates each string as a player (the target) and plots each player who they played on an all-star team with as the source. The relationship is undirected as the dynamic between two players is a two way street: because one player is selected to an all-star team does not mean another one is chosen too (at least in theory, one would hope).

Results

I was pleasantly surprised to see both expected and unexpected results from the Gephi visualization. While some stars have obvious weight that they pull, such as Kevin Garnett and Tim Duncan, the presence of Ray Allen and Kevin Love as central characters sharing teammates from many teams and years is more surprising. What is also interesting is the ways in which the groupings formed, with distribution of stars from the Western Conference all-star teams along the left side of the viz and stars from the Eastern Conference along the right side of the viz. Players who are strictly to one of these sides likely played all their all star games during this period on one conference, whereas those in the center split time between teams on the east and west coasts, and thus had the opportunity to play with all stars from both conferences. For instance, Kevin Love (represented as the top most center circle), has a smaller circle size than Parker, Wade or Bosh on the edges, but has a central placement for his years as an all star both as a Minnesota Timberwolf (Western Conference) and as a Cleveland Cavalier (Eastern Conference).

A very interesting pattern that is made visible from this visualization is the difference among star changes in the Eastern and Western Conferences. Whereas the Western Conference has fewer overall players represented with a larger circle size overall and placement closer together, the Eastern Conference seems to have more smaller groupings of stars within this time frame, with many more one-off stars who gravitate around players who played in a few all star games over a smaller period. While the west has generation-defining stars like Kobe Bryant, Tim Duncan, and Dirk Nowitzki who were selected for many, if not most, of the all star games during this period and collectively played with nearly every Western Conference all star, the Eastern Conference has more isolated periods of star dominance, stretching the right side of the viz further out. Similarly, when one zooms in further to the small pockets of one-off stars on the edges of the viz, they can notice patterns of time changes, with the lower most circles reflecting all stars in the early 2000s, the central outer circles displaying the latter 2000s, and the highest outer circles showing the most recent new all stars. These one-off players gravitate around larger stars who bridged eras together, somewhat like planets. It is also interesting to see which stars pull close together, such as LeBron James, Dwyane Wade, and Chris Bosh who shared many years on the Eastern Conference all star team, as they were drafted the same year to east coast teams, played together for years on the Miami Heat, and all remained in the east their whole careers (besides James, who has played for the Los Angeles Lakers since 2018, which is after this data was collected).

Lastly, it is important to caution viewers that this visualization does not represent who is the biggest star, the best player, or who had the most beloved following in the league. This data reflects who played in the most all star games in a given time frame, who they played with, and if they played in one of the two conferences throughout their tenure or if they switched teams. That should explain why Ray Allen, a consistent but not universally recognized star, has a larger circle than globally known stars like LeBron James and Kobe Bryant. If this visualization were to represent players’ star recognition or other intangibles, they would surely be in the center drawing large amounts of smaller stars around them.

Future Directions

In future studies, I am interested to learn how to make a network visualization that demonstrates changes over time through animation, so a viewer can see how over years star dominance changes – allowing circles to move and change shape as their prominence changes. Similarly, I would like to integrate more information about the players, where a viewer can interact by clicking a circle and finding out more about their career statistics, information about their background, and other intangible information that fans may want to know that could illuminate their role in this visualization overall.