#MakeoverMonday 2021/W15: Fouls Called By NBA Referees

by Henry Mak

Today after a day of Data Viz 101 with Coach Andy, we were tasked with applying what we learnt to this week's Makeover Monday dataset. This week's dataset is all about the fouls called by NBA referees over both the regular season and playoffs.

The field names in the dataset included:

  • Season
  • Season type (Playoffs or regular season)
  • Referee Name
  • Games
  • Total fouls
  • Number of X fouls (replace X with a basketball foul)

This is what the original visualisation looked like, by Owen Phillips:

What I like about it:

  1. Simple to understand
  2. Clean design
  3. Arrows and annotation to guide usage

What could be improved:

  1. Add tooltips to explain each rule for those new to basketball
  2. What exactly Z-Score?
  3. It doesn't have much depth, though I understand that the original visualisation was more of an exploratory tool
  4. White labels could be put in the centre rather than on the red circles
  5. The design could be crafted so it is associated with the NBA more

Honestly, although I've given 3 good points and 5 bad points I think it's a pretty good viz, overall. I'm tempted to say "if it ain't broke then don't fix it" but let's push myself and try and come up with something different.

Creating the viz:

After loading up the dataset to Tableau I immediately plotted the referees against the number of games they have officiated. This was to find the most experienced referee(s). This turned out to be Scott Foster who had almost 300 games, over 20 years(!), of experience officiating in the NBA. After doing a bit of Googling on this guy, I found out he was voted as NBA's (subjectively) official in the NBA in a 2016 survey by the LA times!

This got me thinking... why do the players, head coaches and assistants think this? Is it because he gives the wrong calls? Is it because he is too harsh/lenient? Is it because he is not personable?

During Data Viz 101 we were taught to ask the data 6 simple questions (in this order specifically):

  1. When?
  2. Where?
  3. What?
  4. Who?
  5. How?
  6. Why?

In this case, I landed on number 5. The questions I wanted to answer was:

How many fouls per game does Scott Foster call?

Perhaps this might give me at least some indication as to why he isn't liked.

I did this by making a scatterplot of fouls per game vs number of games officiated.

Lo and behold, my final visualisation:

As you can see, Mr Foster is up there in the top right showing that he makes the most fouls called per game and the most experienced official. But... there was a point on the scatterplot which interested me. Dan Crawford. Even though he had a comparable number of fouls per game, he was voted as the best official.

Interesting stuff. The takeaway point from this is that there are many factors into what makes a 'good' referee which is as expected, I suppose.

If I had more time, I would have added parameters which would allow you to browse for a certain referee and season. I would have also displayed the types of fouls given when you select a referee. However, I wanted to not spend too much time on this overall as I know that time management is a crucial skill at The Data School and a skill that I need to brush up on.

One limitation of this dataset is that Dan Crawford retired in August 2017. This is why he has less games compared to Scott Foster. Another limitation is that although Scott Foster was voted the worst referee, he might actually be a good one objectively but the players just dislike him which affected their vote. Vice versa with Dan Crawford.

In any case, if you'd like to check out my viz then click this link. I hope you learnt more about my thought process into creating this.

Stay tuned to see my week 1 at the DS blog! It'll be weird since this blog post was meant to be my second and not my first but it is what it is :)

Thumbnail photo by Ramiro Pianarosa on Unsplash