Heat maps are the right choice when you need to spot patterns across large volumes of data that would be impossible to read in a table or spreadsheet. They use color intensity to represent values, letting your eye quickly find clusters, outliers, and trends without scanning individual numbers. The format works across wildly different fields, from website design to genomics to geographic analysis, but each use case has its own logic for when a heat map actually helps and when it creates confusion.
Comparing Values Across Two Dimensions
The most common heat map is a grid where rows represent one variable, columns represent another, and color fills each cell based on the value at that intersection. Think of a table showing sales performance across 12 months and 50 products. Reading 600 numbers is tedious. Color-coding those cells so darker shades mean higher values lets you instantly see which products peak in which months.
This matrix format is ideal when both dimensions are categorical or orderable and you want to find relationships between them. Correlation matrices are a classic example: each cell shows how strongly two variables move together, with color intensity increasing as the relationship gets stronger. A correlation of 0.7 or higher typically appears as a deep color (often blue for positive, red for negative), while values below 0.1 fade to white or gray, signaling no meaningful relationship. You can scan dozens of variable pairs in seconds instead of reading a table of coefficients.
Analyzing Website and App Behavior
In user experience research, heat maps solve a specific problem: understanding what visitors actually do on a page versus what you designed them to do. Three types cover the main behaviors.
Click heat maps show exactly where users click and how many clicks each area receives. This reveals whether people are hitting your call-to-action button or clicking on decorative elements they mistake for interactive ones. It also exposes clickable elements that don’t stand out enough to get noticed at all.
Scroll heat maps divide the page into horizontal bands and show the average visibility and time spent on each section. If 80% of visitors never scroll past your third paragraph, anything below that fold is essentially invisible. This is particularly useful for long-form pages, landing pages, and product listings where content order directly affects conversions.
Movement heat maps track where visitors move their cursors. Since mouse position loosely correlates with visual attention, these maps highlight which parts of a page receive the most focus. They’re most useful when combined with click data to distinguish between areas people look at and areas they interact with.
Visualizing Gene Expression and Biological Data
In genomics, heat maps became a standard tool because researchers routinely need to compare expression levels across thousands of genes and dozens of samples simultaneously. Each row represents a gene, each column represents a sample, and color indicates whether that gene is more active (often shown in red), less active (blue), or unchanged (black) compared to a baseline.
The real power comes from pairing the heat map with clustering algorithms that group genes and samples by similarity in their expression patterns. This combination can reveal sets of genes that are commonly regulated together or biological signatures tied to a specific disease or environmental condition. Without the visual layer, those patterns would stay buried in massive data tables. Use a heat map here when you’re looking for broad patterns across conditions, not when you need precise expression values for individual genes.
Mapping Geographic Density
Geographic heat maps overlay color gradients on a map to show where events or features concentrate. Crime incident maps, disease outbreak tracking, and customer location analysis all use this approach. The color represents density: hot spots where many data points cluster appear as intense reds or oranges, while sparse areas fade to cooler tones.
This type of heat map works best when you have a large number of point features, especially when individual markers would overlap and become unreadable. If you’re plotting 10 locations on a city map, pins or dots work fine. If you’re plotting 10,000 incidents, a heat map reveals the spatial pattern that individual points obscure. For sparse distributions, aggregation methods like grouping points into regions or hexagonal bins tend to communicate more clearly than a heat map with isolated blobs of color.
Medical and Thermal Imaging
Infrared thermography generates literal heat maps by capturing surface temperature across the body. Deviations in skin temperature can indicate inflammation, infection, changes in blood flow, or the formation of new blood vessels. The technique became widely familiar during the COVID-19 pandemic, when thermal cameras were used for non-contact body temperature screening in airports and public spaces.
In clinical settings, thermal heat maps are useful for detecting conditions that correlate with temperature changes in underlying tissue, including neurological, vascular, and metabolic issues. The advantage is that the imaging is completely non-invasive and produces an intuitive visual output where hot spots literally look hot.
When a Heat Map Is the Wrong Choice
Heat maps fail in predictable ways. If your dataset has only a handful of data points, a bar chart or scatter plot will communicate more clearly. Heat maps depend on density and pattern to justify their complexity. Five data points in a color grid just look odd.
They also struggle when precision matters more than pattern recognition. Color is inherently imprecise as a visual encoding. Most people can distinguish maybe 5 to 7 shades on a gradient before they start guessing. If your reader needs to know the exact difference between two cells, a labeled table or bar chart is better. Heat maps trade precision for the ability to show the big picture at a glance.
Outliers can distort the entire visualization. A single extreme value compresses the rest of the color scale, making moderate differences invisible. Normalizing your data before generating the map helps prevent this. In scientific contexts, between-sample normalization methods tend to produce more consistent and reliable results than within-sample methods, though the tradeoff is potentially missing some true positives.
Color Choices and Accessibility
Roughly 8% of men and 0.5% of women have some form of color vision deficiency, which means a red-green heat map is unreadable for a significant portion of your audience. Sequential palettes that move from light to dark in a single hue (like light yellow to dark blue) are safer. Diverging palettes that use two distinct hues work well for data with a meaningful midpoint, like positive versus negative correlations, but both hues need to remain distinguishable for color-blind viewers.
Resources like ColorBrewer and Paul Tol’s color schemes provide tested palettes with specific RGB values designed for accessibility. Choosing the right palette isn’t just a courtesy. It determines whether your heat map communicates its data or hides it behind a pretty but illegible gradient.
Choosing the Right Moment
Use a heat map when you’re working with a dense dataset that spans two dimensions and your goal is to find patterns, clusters, or outliers across the whole surface rather than reading individual values. The format shines when the question is “where are the hot spots?” rather than “what is the exact value?” Whether that surface is a website page, a genome, a correlation matrix, or a city map, the logic is the same: too much data for a simple chart, and the interesting story lives in how values cluster together.