The field of data visualization is moving towards more transparent and modular approaches, with a focus on preserving the global geometry of the underlying data. Researchers are exploring new methods for dimensionality reduction, clustering, and embedding, which aim to provide more accurate and informative visualizations. One notable trend is the incorporation of curvature-based metrics to enhance the visualization of high-dimensional data. Another area of interest is the development of optimization techniques for arranging and coloring partitions in multivariate visualizations. Furthermore, there is a growing interest in representing complex datasets as manifolds, which can help illustrate topological characteristics and provide insights into important properties of the data. Noteworthy papers include: EmbedOR, which proposes a stochastic neighbor embedding algorithm that incorporates discrete graph curvature, and Matisse, which describes a methodology and system for generating and visualizing manifolds from Internet latency measurements.