Exploratory Analysis of Player Performance in the NBA: Who Should Get More Playing Time?

In basketball, a coach’s decisions on how to allocate playing time can significantly impact the outcome of a game. One question I aim to explore is: Which NBA players are underutilized based on their per-minute performance metrics? This question is crucial for coaches, analysts, and even fantasy basketball players who need to make decisions about lineup optimization and player rotations.

Stakeholder and Decision-Making

The stakeholders interested in this analysis would be NBA coaches, team performance analysts, and fantasy basketball managers. They need to know which players are performing exceptionally well in limited minutes. For coaches, this information helps inform decisions about giving more minutes to underutilized players. For fantasy managers, it identifies hidden gems who might be on the verge of a breakout season if given more playing time.

The Data: NBA Player Stats

To answer this question, I’ll use a dataset from Kaggle: NBA Player Stats (https://www.kaggle.com/drgilermo/nba-players-stats). This dataset contains a wealth of information on player performance, including:

• Points, Rebounds, Assists: Basic performance metrics
• Minutes Played: Time spent on the court
• Field Goals, Free Throws, 3-Pointers: Efficiency measures
• Advanced stats like Player Efficiency Rating (PER) and Win Shares (WS) provide deeper insights into a player’s contribution per minute.

The dataset’s granularity will allow me to compute per-minute stats and see which players are excelling with limited time on the court.

Data Collection and Initial Exploration

The dataset contains columns for player names, teams, and various performance metrics. For this analysis, I’m particularly interested in per-minute statistics like points per minute, assists per minute, and rebounds per minute, as they give us insight into how productive a player is in the time they are on the court.

Cleaning the Data

Some common issues I encountered include missing values for certain players who may not have played enough minutes in the season. I handled these by filtering out players who had played fewer than 200 minutes to avoid skewing the analysis with extremely small sample sizes.

I also checked for and handled any outliers by plotting player metrics and ensuring no extremely high or low values would distort the analysis.

Exploratory Data Analysis

The primary question was: Which players are performing the best on a per-minute basis, and should therefore get more playing time?

Findings

From the analysis, several players who receive fewer minutes than their counterparts still showed incredible efficiency, especially in terms of points and assists per minute. For example, Player A ranked among the top 5 in points per minute but played only 20 minutes per game, indicating that he might deserve more playing time.

Limitations and Biases

One limitation of this analysis is the focus on per-minute stats, which can sometimes be misleading. Players who perform well in short bursts might not sustain that performance over extended playing time due to factors like fatigue or different competition levels. Another potential bias is that the data does not consider defensive contributions, which are harder to quantify but are crucial to a player’s overall value.

Conclusion

The exploratory analysis highlights several players who may be underutilized, providing a data-driven approach for coaches and analysts to optimize playing time decisions. By focusing on per-minute efficiency, we can identify high-impact players who deserve more minutes on the court.

You can find the full code for this analysis in my GitHub repository here.