Game Changers: Unravelling NBA Player of the Week Insights
In my first SQL project for my portfolio, I take a look at the past winners of the NBAs Player of the Week award. This dataset was provided by Khan Academy, as part of their SQL course which I recently completed. I carried out the analysis using the online SQL editor that Khan Academy provides on its website.
The Data
The dataset originally included all winners dating back to October 1979 but I decided to reduce the dataset, partly because the online SQL editor I used was struggling with such a big dataset and also because I’m personally more interested in the recent players as I’ve watched them play!
So I filtered the dataset to only show the winners from 2005–2020. With seasons running for 6 months each year and two winners for each week of the season, one for each conference, this left me with 690 rows of data.
Interesting columns in the dataset included:
- Player
- Team
- Conference
- Position
- Height
Before getting started, I outlined a few questions which I wanted to answer as starting points:
- How many awards have been given out in this period and how many different players have won these?
- Of these players, who has won the award the most? Which players make up the top 10?
- What is the average height of the winners? What does the general height distribution of winners look like?
- What is the average height across positions? Which position has the tallest players?
- Who are the top 10 tallest winners? Do they all play in the position from Q4?
The dataset is available to download on Kaggle here.
Summary
Before diving straight into the analysis, here is a high-level summary outlining the answers to the above questions which I found:
- The 690 Player of the Week titles over this period have been shared amongst 154 players.
- Perhaps as expected, Lebron James is the most decorated player when it comes to Player of the Week awards, with his 62 wins putting him comfortably ahead of second-placed Kevin Durant, who got 26.
- The average height of the winners is 200cm, with the height distribution being discussed in more depth later.
- The tallest players are usually centers, with an average height of 213cm.
- 90% of the 10 tallest players to win this award play center.
I also went down a couple of rabbit holes and added in some additional analysis in certain sections.
The Analysis
To find out how many awards have been given out in this period, a simple COUNT(*) function was enough, answering 690. Adding in a DISTINCT statement, we can see that these 690 awards were won by just 154 players.
Following on from this, I wanted to see which 10 players had won the award the most. To do this I used the following commands:
This showed, unsurprisingly, that Lebron James has been crowned the Player of the Week more than any other player between 2005 and 2020. His 62 wins put him comfortably ahead of second-placed Kevin Durant, who only won the award 26 times over the same period.
Now I wanted to take a look at the height distribution of all the winners. To start, I determined that the average height of all winners was 200cm. (This is compared to the average American male who is around 175cm tall!) After this, I divided players into height categories using the CASE function, to see how many of the winners have been above/below or equal to average height.
This showed that of the 690 awards given out, there has never been a winner who was exactly 200cm tall. Further, the number of winners who have been above average height (403) is 116 more than the number of winners who have been below average height (287). This indicates that most of the winners who are above average are only slightly above average and those who are below average can be well below the average.
This makes sense as these NBA players are close to being as tall as you can physically get, so there is not much room for high-value outliers, while there is definitely scope for some (relatively) shorter players to make it. Think Allen Iverson (183cm) or Isaiah Thomas (175cm), who won the Player of the Week award 10 and 5 times respectively.
Let’s look at the quartile values to get a better idea of the distribution. As there are no actual functions to calculate medians and quartile values in SQL this needs to be done manually. I combined the LIMIT and OFFSET functions to calculate the upper quartile, lower quartile, and median.
This gave me the below results:
These results support my earlier statement, with the median being above average (203cm) and the upper quartile (208cm) being only 5cm higher than this. The lower quartile, meanwhile, is 13cm below the median, coming in at 190cm.
Looking at the different positions, I wanted to get an idea of the average height by position. To do this I used the GROUP BY function as shown below.
This shows that, on average, centers are the tallest players, with an average height of 213cm. This means that an average center is 27cm taller than an average point guard, who are the shortest players on average coming in at 186cm.
After this, I wanted to know if the top 10 tallest players in the data set were all centers, as the above result would suggest. To see if this was the case, I used the ORDER BY and LIMIT functions. This one gave me a bit of trouble as I initially forgot to include the DISTINCT statement and was left with a repetitive list of “Yao Ming” but I eventually realized my mistake and was left with the below.
8 of the 10 tallest Players of the Week from 2005 to 2020 were centers, with a 9th being a forward-center, who switches between forward and center. Nikola Vucevic, a power forward, is the only player who never plays center on the list.
A Final Word
Thank you so much for making it this far and reading my entire Game Changers NBA project! As I mentioned above, this is my first SQL project. I hope you enjoyed reading it as much as I enjoyed putting it together. Whether or not you enjoyed it, I’d love to hear your feedback so please feel free to get in touch — you can reach me on LinkedIn here or by email at jamesonjoshua1@gmail.com.