Data Study NBA Stats (1947-2015)
The skills the author demoed here can be learned through taking Data Science with Machine Learning bootcamp with NYC Data Science Academy.
Introduction
This has been an exciting season in the NBA. Throughout the season, the headlines on the NBA site read βWarriors having the best start to a seasonβ, βLebron James becoming the youngest player to hit the 26,000 points milestoneβ, βRussell Westbrook having the most triple-doubles in the past 50 yearsβ, βSteph Curry shooting over 400 3-pointersβ, and of course the βGolden State Warriors obtaining the best winning ratio of a seasonβ. And there were many more. This lead me to create a Shiny app see what to other data records are there that can be broken in the future.
Data Set
I located a dataset from Putdat that had NBA stats from 1947 β 2015. The dataset contained all the teams that ever existed from 1947 even though the team was discontinued. However the stats of the teams that changed names or changed location were preserved under the new name.
Data
Stats Leader
The first tab of the app is the Stats Leaders. This section is where we can find the overall stats of the NBA. The graph below shows the teams that won at least one championship. The Boston Celtics won an astounding 17 championships (eleven of them was during Bill Russellβs career). So far 17 out of the 30 teams in the current NBA won at least one championship. That leaves 13 teams to win their first NBA championship and break the streak.
Win
The Wins graph (below) shows the top ten records of teams from 1947 through 2015 in both the seasons and the playoffs. The first graph shows the top ratios of wins to total games and using the slider, we can adjust the range from one to ten (five is shown in the screenshot). For instance (before the Golden State Warriors stunning record of 73-9), the best record was the Chicago Bulls record of 72 wins and 10 losses for a winning percentage of 88% in 1996. The Bulls also managed to maintain the best winning ratio in the season for the following year.
Average PPG
On the next tab, Team, team stats are also available for both the season and the playoffs. Sticking with our Bullsβ example, we get these two graphs:
The first graph (red bar graph) shows the points per game every year since the start of their franchise. It is interesting to see that the Bullsβ best record in β96 was not their highest point performance. This can apply that they had a great balance between offense and defense that helped them win that many games.
The second graph shows the winning ratio split into home and way games. This is the Bullsβ season winning ratio. One last point I want to make about Bulls is that after the Michael Jordan era, the Bullsβ faced a sharp decline in points and wins in general. It was not until the signing of Derrick Rose that some light started shining in the windy city. It is actually interesting to see that majority of the teams win more home games than away. This just proves that fans are part of the team. There are definitely a few exceptions. The Timberwolves had an away winning percentage of 32% and their home was 27%.
Conclusion
With more time, I will be able to create more graphs that demonstrate more current records. Also it was interesting to see trends through the graphs. When I saw them, I wanted to know why this was the case or which player made an impact that year. For instance I saw that Bulls had a phenomenal record in the year 1996. Not only would I want to be able to click on the bar for 1996 and see Jordan, Pippen, and Rodman along with their respective stats. I would also like to find a dataset with individual stats and include them into this Shiny app.