Fig 1.
Radar plots of positional performance profiles based on raw and normalized metrics.
Each panel represents one of the five basketball playing positions (Point Guard, Shooting Guard, Small Forward, Power Forward, and Center). Orange lines display the average raw metric values, while red lines represent the normalized values used in subsequent analyses. Metrics include various scoring, rebounding, passing, and defensive indicators.
Table 1.
Differences in raw and normalized per-40-minute performance metrics across traditional playing positions.
Fig 2.
Relationship between raw mean values and normalized values of performance metrics across playing positions.
Each point represents a specific basketball performance indicator, with shape indicating the metric and color representing the player’s position. The x-axis shows the raw mean value of each indicator, while the y-axis displays its normalized (standardized) value used for clustering and classification analysis. The dashed diagonal line represents equality (i.e., points where the normalized value equals the raw mean), serving as a visual reference for relative scaling across metrics.
Fig 3.
Optimal number of clusters according to average silhouette width.
Fig 4.
Distribution of playing positions across the nine identified clusters.
Stacked bar chart showing the number of players in each cluster, broken down by specific playing positions (Point Guard, Shooting Guard, Small Forward, Power Forward, and Center). The height of each bar represents the total number of players assigned to a given cluster, while the color segments within each bar indicate the positional composition.
Table 2.
Differences in raw performance metrics across the 9 clusters.
Fig 5.
Comparison between mean and normalized values of performance indicators across nine player clusters.
This scatter plot displays the relationship between raw mean values (x-axis) and normalized values (y-axis) for each performance metric, with color indicating the metric type and shape representing the player’s assigned cluster (1–9). The dashed diagonal line denotes equality between the mean and normalized scales. Markers above the line indicate metrics where normalization disproportionately amplified values, while those below indicates compressed scaling.
Fig 6.
Radar plots of the nine functional player clusters based on raw and normalized performance metrics.
Each panel represents a distinct cluster identified through unsupervised classification of performance indicators. Radar plots show the average raw (orange line) and normalized (red line) values for each metric, including scoring, rebounds, assists, turnovers, steals, and blocks. The average playing time for each cluster is indicated below the plots.
Fig 7.
Heatmap of raw players’ performance classification into 9 clusters.
Color intensity represents the relative importance of each performance indicator within its row, normalized from 0 (red, lowest) to 1 (green, highest). Data are means.
Fig 8.
Heatmap of normalized players’ performance (per-40-minute) classification into 9 clusters.
Color intensity represents the relative importance of each performance indicator within its row, normalized from 0 (red, lowest) to 1 (green, highest). Data are means.
Table 3.
Summary table outlining the nine functional profiles, described by tactical characteristics, representative performance indicators, and in-game roles.