Density Score

The density_score quantifies the local data density surrounding each data point within the 2D UMAP embedding space. It is calculated by applying a Kernel Density Estimation (KDE) using a Gaussian kernel on the two-dimensional embeddings of the dataset. The bandwidth parameter of the KDE (defaulted to 0.1 but tunable) controls the smoothness of the density estimation, influencing how local or global the density estimation is. The resulting density_score reflects the density distribution of data points in the 2D space.

Note

To further assist users with data curation tasks, Kolena automatically calculated a number of metrics based on the embedding space details. Enable automatic embedding extractions or upload your own embeddings to utilize these scores.

Interpretation

A higher density_score indicates that a data point is located in a region with many neighboring points, suggesting that it represents a common pattern or feature within the dataset. Conversely, a lower density_score suggests that the data point resides in a sparsely populated region, potentially highlighting outliers or rare patterns. This metric is particularly useful for visual data exploration, enabling you to identify clusters, anomalies, and understand the overall structure of your dataset in the embeddings visualizer.