Single-cell transcriptomic datasets have enabled the study of gene expression at an unprecedented resolution and scale. The high-dimensional and large-scale nature of single-cell transcriptomic landscapes necessitate efficient and accurate computational tools for extracting biological insights from these data. Unfortunately, standard analysis workflows often neglect information about local density in the original transcriptomic space, resulting in misleading representations of transcriptomic variability of individual cell states in downstream analyses. In this talk, I will introduce our recent algorithms for single-cell analysis that expressly account for density differences in the underlying dataset: densMAP and denSNE for data visualization, which respectively augment widely-used methods UMAP and t-SNE, and GeoSketch for sketching (downsampling) massive datasets. Our methods facilitate more accurate and unbiased exploration of single-cell transcriptomic landscapes.