Finding the Right Fit: An NBA Analysis using Tableau

The Data

Being a former professional basketball player, I thought analyzing data from the 21-22 NBA season would be a TON of fun.?I was not disappointed!? Without question, basketball, especially the NBA, has worldwide popularity due to a multitude of factors; celebrity athletes, worldwide exposure, and fantastic branding.? I enjoyed using my "basketball brain" to find the answers I came up with.

My role in this project is that of an analyst with the Detroit Pistons.?The general manager, and assistant coach have questions that need to be answered quickly and I used Tableau to create visuals to answer those questions.

The dataset is from a basketball reference website for the 2021-2022 season and encompasses points, assists, rebounds, turnovers etc. for over 600 players that played on one of the NBA’s 30 teams.

You can view the dataset here:?2021-2022 NBA Data Set

I worked to answer the following questions:

?

·????????Who are potential free agent signings that we can target based on last

season’s stats?

·????????What position is the MOST effective at 3-pt shooting?

·????????Which players were the top scorers on each team?

·????????What players had the most assists at every position?

·????????My direct leader wants to know: what have I been working on

lately?

Insights

1.??????I can see from the bubble plot that SG’s and PG’s have higher points and assists because they handled the ball, therefore, they have more of a direct impact on the plays. ?Nikola Jokic is an outlier as he is a center that scores a lot, but also has a higher number of assists for his position.

2.??????From the heat map I created, we can see the PF position has the highest 3-pt percentage on average out of the 30 teams, followed by the SF, SG, PG and C. This is very surprising as most PF's are big guys who are working inside as opposed to shooting 3's.

3.??????Looking at the stacked bar chart we see that the top few scorers are Giannis Antetokounmpo, DeMar DeRozan, Joel Embiid, and Trae Young.

4.??????My original analysis showed that James Harden had the most assists in the NBA in the 21-22 season with 1,334. We will need to look into this further as this is a VERY high total.?????????????????????

5.??????I have put together all this data into a story format on Tableau for easier use and sharing with leadership and the entire analytics division.

Analysis

?

1.?????Who are potential free agent signings that we can target based on last season’s stats?

For this portion, we are being asked to get a feel for how different players performed last season to evaluate for potential free agent signings.?I am trying to see how different players performed looking at their total points, assists and rebounds.

First, I used 3 quantitative variables of points, assists and rebounds along with the qualitative variable of positions.?Then, by dragging the "position" into the color shelf, the "player" into the text shelf and the sum of total rebounds to the size shelf I was able to create this bubble plot.

The bubble plot shows every player, and the size of their circle is based on rebounds.?The main pattern I see here is that the SG’s and PG’s, in the green show having higher points and assists totals. ?The one major outlier here is Nikola Jokic, who is a center, showing over 2,000 points and just under 600 assists.

looking through the data, I think that recommending a few players to the GM for possible free agency pick-ups would be D’Angelo Russell, Jrue Holiday or Kyle Lowry

2. ??What position is the MOST effective at 3-pt shooting?

?

The assistant coach of the Pistons wants us to create an easy way to know which position is the most efficient at 3-pt shooting based on percentage for every team.

?To accomplish this, I place the "Teams" in the rows and "Position" in the columns section.? Next, we take the "3P%" and add that to the text shelf then change the metric from SUM to AVG.?Then, to give the heat map a little color we add the “3P%” to the color shelf and change the marks section from “automatic” to “square”.

Now these patterns are extremely easy to read, especially during the high intensity of a game.?We can see the higher percentages are in dark blue for each team.

3. ????Which players were the top scorers on each team?

To find a creative way to show which scorers were the top for each team this year I decided to go with a stacked bar chart.?I started off by creating a simple bar chart in Tableau.?I showed the "Teams" vs. "Total points" scored.

This is far too boring, and we still need to show who scored the points, so I added the “player” column to the label shelf and the color shelf.? Finally, to make this a little bit easier to read I right clicked on the axis name of “Tm” to make my final stacked bar chart.

We can see that some of the highest scorers for each team were Giannis Antetokounmpo, DeMar DeRozan, Joel Embiid, and Trae Young.

4.??????What players had the most assists at every position?

?

The next request was from the GM of the Pistons, and they want to improve the assist numbers for next season.?They would like to know what players had the most assists in every position.

Using the “AST” we add that to the size shelf, then take “Pos” and put that in the color tab.? Basically, at this point the chart looks like a large color mosaic.?Then, we then dragged the position and assists to the text shelf, and we now have the following:

The last thing to do is add the player name to the text shelf and we have our finalized tree map:

Then, I thought I was done… However, being the basketball nerd I am, the assists for James Harden looked incredibly high for a player who only played 65 games that season.?If we do the math, it comes out to ~20 assists PER GAME!?This is an astronomical number of assists, especially if you know how James Harden plays (He shoots... A LOT).? I dug into the data further and found that the data being used was not taking into account that he was traded, essentially, doubling his assist total:

The dataset shows that James Harden’s total assists were 667 and he played 65 games that season giving him an average of 10.2 assists per game.?The extra rows added in his assist total doubled due to him being traded mid-season. ?To fix this, I cleaned the dataset to get rid of any duplicates to make sure we had accurate and clean numbers for my general manager (something I should have done prior to creating all of my charts).

5.???????The last business question my leader wanted to know was:?

What have you been working on lately?

I showed my team leader that I took all the data and put it into these graphs and charts, but I took it a step further and created a story in Tableau.? This story allows me to give a little insight for each chart and it allows others to see the interactive side of the document.? With this story mode, I can send the Tableau link to my team, and they can look at specific players to see who we may want to target, as an organization, and give ourselves the best chance of finding the correct player (statistically) for the Detroit Pistons to succeed.

?

I have created a dashboard story using all of my previous graphs and charts to bring everything together.

Recommendations

?

-?????????I think that recommending a few players to the GM for possible free agency pick ups would be D’Angelo Russell, Jrue Holiday or Kyle Lowry.? These players are all high scorers and PG’s that have a great level of assist to turnover ratios. ?Focusing on these players could give a boost to overall scoring for the team, which is a struggle for the current Pistons team.?

-?????????In speaking with our assistant coach, I see that our team’s 3-pt shooting by position is well under the average for the league. ?I would recommend finding a guard who has more of a shooting touch, a combo guard, or a pure shooting guard to help drive our offense.

-?????????Lastly, looking at bringing in a player(s) who can drive up our assist numbers I would look at young guys who have relatively smaller contracts who are hungry for playing time OR older more established players who can bring our younger team experience. ?Since Detroit is in the middle of a rebuild, I would focus on three players specifically to fill this role:

o??Caris LaVert

o??Mike Conley

o??Denzel Valentine

Thank you, as always, for reading through my project, I hope you found it as fun to read as I did to create! I am open to any and all feedback regarding this article and suggestions on how to improve!

Check out this project and my other projects on the featured section of my profile!

For my Tableau Dashboard, click below.

Bubble | Tableau Public