A data heatmap is a color-coded grid for a two-dimensional numeric matrix. Rows and columns remain discrete categories, while the color scale makes high, low, similar, and unusual patterns visible without reading every number. It is useful for gene expression, correlation, classification, activity, and business data.
Sequential palettes suit values that progress from low to high. Diverging palettes suit values around a meaningful midpoint such as zero, change from baseline, or a Z-score. The legend is part of the evidence: it must state what the colors mean.
Clustering can reorder rows or columns so similar profiles sit together. A dendrogram records that ordering, but it does not prove biological or causal groups. Interpret clusters alongside the study design and source data.
Matrix
Rows and columns define the observations or variables being compared.
Color scale
A legend maps numeric values to color and identifies the midpoint.
Normalization
Row or column Z-scores change the question from magnitude to relative pattern.
Clustering
A distance metric and linkage rule reorder similar profiles together.









