Heat Map Examples: 7 Types of Heat Maps with Real Use Cases

"Heat map" is really an umbrella term for seven different charts that share one idea: encode a number as color intensity on a grid or surface. The right one depends on what your rows and columns represent — variables, dates, hours, regions, pixels, or cohorts.

If you need the fundamentals first — what the color scale encodes and when a heat map beats a table — start with heat maps explained. Below are the seven types you'll actually build, each with what it shows, when to use it, and the fastest tool for the job.

1. Correlation Matrix Heat Map

What it shows: pairwise correlation coefficients between every numeric variable in a dataset, colored from -1 (strong negative, usually blue) through 0 (white) to +1 (strong positive, usually red). Clusters of dark cells reveal groups of variables that move together.

When to use it: the first ten minutes of exploratory data analysis, feature selection before modeling, or checking for multicollinearity in a regression.

Tool: Python's seaborn does this in two lines — see the full seaborn heatmap tutorial for the mask trick that hides the redundant upper triangle:

import seaborn as sns
sns.heatmap(df.corr(), annot=True, cmap="coolwarm", center=0)

2. Calendar Heat Map (GitHub-Style)

What it shows: one colored square per day, arranged in a week-by-week grid across a year. Darker squares mean more of something — commits, sales, workouts, support tickets. GitHub's contribution graph made this layout famous.

When to use it: daily count data spanning months, where you care about streaks, gaps, and weekly rhythm ("why are Sundays always empty?"). A line chart of the same data hides the weekday pattern; the calendar layout exposes it.

Tool: calplot or july in Python, ECharts' calendar coordinate system in JavaScript, or a clever Excel grid (row per weekday, column per week) with conditional formatting.

3. Time-of-Day × Weekday Traffic Heat Map

What it shows: a 7 × 24 grid — weekdays down the side, hours across the top — with color showing volume: website sessions, call-center load, gym check-ins, emergency room arrivals. The "hot" blocks jump out instantly.

When to use it: staffing and scheduling decisions. This is the single most actionable heat map in business analytics because the answer ("Tuesday 10am–1pm is peak") is directly operational.

Tool: a pivot table (weekday rows, hour columns, count values) plus a color scale — in pandas + seaborn, Excel, or any BI tool. Build one step-by-step in how to make a heatmap.

pivot = df.pivot_table(index="weekday", columns="hour",
                       values="sessions", aggfunc="sum")
sns.heatmap(pivot, cmap="YlOrRd")

4. Geographic Heat Map (Choropleth)

What it shows: regions on a map — countries, states, zip codes — shaded by a value: population density, sales per region, election margins, COVID case rates.

When to use it: whenever "where?" is the question. One caution: large, sparsely populated regions dominate visually, so prefer rates (sales per capita) over raw totals, or the map just shows you where people live.

Tool: Plotly's px.choropleth, Folium, Tableau, or Power BI's built-in map visuals. For point data (crime incidents, delivery drops) use a density heat map layer instead of region shading.

5. Website Click / Attention Heat Map

What it shows: a screenshot of a web page overlaid with color showing where users clicked, moved the mouse, or how far they scrolled. Red blobs = heavy interaction; cold areas = ignored content.

When to use it: UX and conversion optimization — finding buttons nobody clicks, non-clickable elements everyone clicks, and the scroll depth where readers give up.

Tool: this one you buy rather than build: Hotjar, Microsoft Clarity (free), or Crazy Egg record real user sessions and render the overlay for you.

6. Cohort Retention Heat Map

What it shows: rows are signup cohorts (e.g., "January signups"), columns are periods since signup (month 0, 1, 2, …), and each cell is the percentage of that cohort still active. Color makes the retention cliff — and any improvement in newer cohorts — visible at a glance.

When to use it: subscription and product analytics. Reading down a column compares cohorts at the same age; a triangle that gets darker toward the bottom rows means retention is improving over time.

Tool: pandas pivot_table + seaborn, or built-in cohort reports in Amplitude and Mixpanel.

retention = df.pivot_table(index="cohort_month",
                           columns="months_since_signup",
                           values="pct_retained")
sns.heatmap(retention, annot=True, fmt=".0%", cmap="Greens")

7. Risk Matrix Heat Map

What it shows: a small grid — typically 5 × 5 — with likelihood on one axis, impact on the other, and cells colored green → yellow → red. Individual risks are plotted into cells so stakeholders see what needs attention first.

When to use it: project risk registers, security assessments, compliance reviews. Its power is communication, not precision: the categories are judgment calls, but the red corner focuses a meeting instantly.

Tool: Excel or PowerPoint is genuinely fine here — a 5 × 5 range with fixed cell colors and labels typed in. No library needed.

Which Heat Map Example Should You Use for Your Data?

Match the grid axes to your question. Variables × variables → correlation matrix. Days across a year → calendar. Hours × weekdays → traffic grid. Regions → choropleth. Page coordinates → click map. Cohorts × age → retention grid. Likelihood × impact → risk matrix. If your data doesn't naturally form a grid, a heat map is probably the wrong chart — reach for a bar or line chart instead.

Question	Heat map type	Fastest tool
Which variables move together?	Correlation matrix	seaborn
What does daily activity look like over a year?	Calendar heatmap	calplot / ECharts
When are we busiest?	Hour × weekday grid	Pivot + color scale
Where is the value concentrated geographically?	Choropleth	Plotly / Tableau
Where do users click?	Click map	Clarity / Hotjar
Are newer customers sticking around longer?	Cohort retention grid	pandas + seaborn
Which risks matter most?	Risk matrix	Excel