Uniform Manifold Approximation and Projection is a more recent algorithm for visualization and general dimensionality reduction. It assumes the data is uniformly distributed on a locally-connected manifold and looks for the closest low-dimensional equivalent using fuzzy topology. It uses a neighbors parameter that impacts the result similarly as perplexity above.

It is faster, and hence scales better to large datasets than t-SNE, and sometimes preserves global structure than better than t-SNE. It can also work with different distance functions, including, for example, cosine similarity, which is used to measure the distance between word count vectors.

The four charts in the bottom row of the previous figure illustrates how UMAP does indeed move the different clusters further apart, whereas t-SNE provides more granular insight into the local structure.

The notebook also contains interactive Plotly visualizations for each algorithm, which permit the exploration of the labels and identify which objects are placed close to each other.