Colormaps (sequential, diverging, qualitative) in Matplotlib - Deep Dive

Overview - Colormaps (sequential, diverging, qualitative)

What is it?

Colormaps are sets of colors used to represent data visually, especially in charts and images. They help show differences or patterns by mapping numbers to colors. There are three main types: sequential, diverging, and qualitative, each suited for different kinds of data. Using the right colormap makes data easier to understand at a glance.

Why it matters

Without colormaps, data visualizations would be dull and confusing, making it hard to spot trends or compare values. Good colormaps highlight important differences and avoid misleading the viewer. They make complex data accessible and help people make better decisions based on what they see.

Where it fits

Before learning colormaps, you should understand basic plotting and color theory. After mastering colormaps, you can explore advanced visualization techniques like color normalization, perceptual uniformity, and interactive plotting.

Mental Model

Core Idea

Colormaps translate numbers into colors in a way that matches the data's meaning and helps people see patterns clearly.

Think of it like...

Choosing a colormap is like picking the right paint colors for a map: some colors show a smooth rise in height, others highlight where a river splits, and some just label different countries with distinct colors.

┌───────────────────────────────┐
│          Colormaps            │
├─────────────┬───────────────┤
│ Sequential  │ Colors change │
│             │ smoothly from │
│             │ light to dark │
├─────────────┼───────────────┤
│ Diverging   │ Colors change │
│             │ from two ends │
│             │ to middle     │
├─────────────┼───────────────┤
│ Qualitative │ Distinct      │
│             │ colors for    │
│             │ categories    │
└─────────────┴───────────────┘

Build-Up - 7 Steps

1

Foundation

What is a Colormap?

Concept: Introduce the basic idea of mapping numbers to colors.

A colormap is a list or range of colors that represent data values. For example, low numbers might be light blue, and high numbers dark blue. This helps us see data differences visually instead of just numbers.

Result

You understand that colormaps turn numbers into colors to make data easier to see.

Understanding that colors can represent numbers is the first step to making meaningful visualizations.

2

Foundation

Types of Colormaps Overview

Concept: Learn the three main types of colormaps and their basic uses.

There are three main colormap types: - Sequential: colors change smoothly from light to dark, good for ordered data. - Diverging: colors change from two different colors to a middle color, good for data with a center point. - Qualitative: distinct colors for categories, no order implied. Each type fits different data stories.

Result

You can name and recognize the three colormap types and their general purpose.

Knowing colormap types helps you pick the right one to match your data's story.

3

Intermediate

Using Sequential Colormaps

🤔Before reading on: do you think sequential colormaps are best for data with negative and positive values or only positive values? Commit to your answer.

Concept: Sequential colormaps are best for data that goes from low to high without a meaningful center.

Sequential colormaps use colors that get darker or more intense as values increase. For example, light yellow to dark red. They work well for data like temperature or population where values only increase or decrease.

Result

You can apply sequential colormaps to show smooth changes in data magnitude.

Understanding when to use sequential colormaps prevents confusion when data has no natural midpoint.

4

Intermediate

Understanding Diverging Colormaps

🤔Before reading on: do you think diverging colormaps highlight differences around a midpoint or just extremes? Commit to your answer.

Concept: Diverging colormaps highlight data that deviates around a central value, like zero or average.

Diverging colormaps use two contrasting colors that meet in the middle at a neutral color. For example, blue to white to red. This helps show values above and below a center, like profit and loss or temperature differences from average.

Result

You can use diverging colormaps to emphasize how data splits around a key point.

Knowing diverging colormaps helps reveal meaningful differences around a baseline.

5

Intermediate

Applying Qualitative Colormaps

🤔Before reading on: do you think qualitative colormaps imply order or just categories? Commit to your answer.

Concept: Qualitative colormaps assign distinct colors to categories without implying order.

Qualitative colormaps use colors that are clearly different from each other, like red, green, blue, and yellow. They are perfect for labeling groups like types of fruit or countries where colors just identify categories.

Result

You can pick qualitative colormaps to clearly separate categories in your data.

Recognizing qualitative colormaps prevents misinterpretation of categorical data as ordered.

6

Advanced

Perceptual Uniformity in Colormaps

🤔Before reading on: do you think all colormaps show equal color differences for equal data differences? Commit to your answer.

Concept: Perceptual uniformity means color changes match how humans see differences evenly across the scale.

Some colormaps change color unevenly, making some data differences look bigger or smaller than they are. Perceptually uniform colormaps, like 'viridis' in matplotlib, ensure equal steps in data look like equal steps in color brightness or hue.

Result

You can choose colormaps that represent data differences fairly and avoid misleading visuals.

Understanding perceptual uniformity improves the accuracy and fairness of data interpretation.

7

Expert

Customizing and Combining Colormaps

🤔Before reading on: do you think combining colormaps can create better visuals or just confuse viewers? Commit to your answer.

Concept: Advanced users can create custom colormaps or combine types to fit unique data needs.

Matplotlib allows blending colormaps or creating new ones by mixing colors or adjusting ranges. For example, combining sequential and diverging colormaps can highlight complex data patterns. But misuse can confuse viewers if not done carefully.

Result

You can tailor colormaps to your data and audience for clearer insights.

Knowing how to customize colormaps unlocks powerful, precise visual storytelling but requires careful design to avoid confusion.

Under the Hood

Colormaps work by mapping numeric data values to colors through interpolation. Internally, matplotlib stores colormaps as arrays of colors in a color space. When plotting, data values are normalized to a 0-1 range, then matched to colors by interpolating between defined color points. This process happens efficiently in rendering pipelines to produce smooth gradients or distinct colors.

Why designed this way?

Colormaps were designed to translate complex numeric data into visual cues humans can quickly understand. Early designs used simple linear gradients, but perceptual studies showed that human vision perceives color differences unevenly. This led to perceptually uniform colormaps and distinct types to suit different data stories. Alternatives like random colors or fixed palettes were rejected because they confuse or mislead viewers.

Data values (numeric) ──▶ Normalize (0 to 1) ──▶ Colormap array (color stops) ──▶ Interpolation ──▶ Color output

┌───────────────┐      ┌───────────────┐      ┌───────────────┐      ┌───────────────┐
│ Numeric Data  │─────▶│ Normalization │─────▶│ Color Lookup  │─────▶│ Final Color   │
└───────────────┘      └───────────────┘      └───────────────┘      └───────────────┘

Myth Busters - 3 Common Misconceptions

Quick: Do you think all colormaps are equally good for any data type? Commit yes or no.

Common Belief: Any colormap can be used for any kind of data without problems.

Reality: Different colormaps suit different data types; using the wrong one can mislead or confuse viewers.

Why it matters: Using a qualitative colormap for ordered data can hide trends, while using a sequential colormap for categories can imply false order.

Quick: Do you think brighter colors always mean higher values in colormaps? Commit yes or no.

Common Belief: Brighter colors always represent higher data values.

Reality: Brightness does not always increase with data value; some colormaps use hue changes or diverging colors where brightness varies non-linearly.

Why it matters: Assuming brightness means higher value can cause misreading of data, especially with diverging or qualitative colormaps.

Quick: Do you think the default matplotlib colormap 'jet' is a good choice for all plots? Commit yes or no.

Common Belief: 'Jet' colormap is a good default for all visualizations because it looks colorful and clear.

Reality: 'Jet' is not perceptually uniform and can create artificial boundaries, misleading interpretation of data gradients.

Why it matters: Using 'jet' can cause viewers to see false patterns or ignore subtle data changes.

Expert Zone

1

Some colormaps are designed to be colorblind-friendly, which is crucial for accessibility but often overlooked.

2

The choice of color space (like RGB vs. LAB) for interpolation affects how smooth and natural the color transitions appear.

3

Combining colormaps with data normalization techniques can reveal hidden patterns but requires careful tuning to avoid distortion.

When NOT to use

Avoid using qualitative colormaps for numeric data where order matters; instead, use sequential or diverging. For data with no meaningful midpoint, diverging colormaps can confuse viewers. When precise perception of differences is critical, avoid non-uniform colormaps like 'jet' and prefer perceptually uniform ones like 'viridis'.

Production Patterns

Professionals often use perceptually uniform sequential colormaps for heatmaps and diverging colormaps for anomaly detection. Qualitative colormaps are common in categorical plots like bar charts or cluster labels. Custom colormaps are created for branding or specific domain needs, and accessibility considerations are integrated into production pipelines.

Connections

Human Color Perception

Colormaps are designed based on how humans perceive color differences.

Understanding human vision helps create colormaps that communicate data clearly without misleading.

Data Normalization

Colormaps rely on normalized data values to map colors correctly.

Knowing normalization ensures you prepare data properly before applying colormaps for accurate visualization.

Cartography

Both colormaps and map colors use color schemes to represent information meaningfully.

Studying map color design reveals principles of effective color use that apply to data visualization colormaps.

Common Pitfalls

#1 Using a sequential colormap for categorical data.

Wrong approach:

plt.scatter(x, y, c=categories, cmap='viridis') # categories are labels, but viridis is sequential

Correct approach:

plt.scatter(x, y, c=categories, cmap='tab10') # tab10 is qualitative for categories

Root cause:

Confusing data type leads to choosing a colormap that implies order where none exists.

#2 Applying diverging colormap without centering data.

Wrong approach:

plt.imshow(data, cmap='coolwarm') # data not centered around zero

Correct approach:

plt.imshow(data, cmap='coolwarm', norm=TwoSlopeNorm(vcenter=0)) # center data at zero

Root cause:

Not normalizing data around the midpoint causes misleading color mapping.

#3 Using 'jet' colormap for scientific data.

Wrong approach:

plt.imshow(data, cmap='jet') # jet is default but not perceptually uniform

Correct approach:

plt.imshow(data, cmap='viridis') # viridis is perceptually uniform and better for interpretation

Root cause:

Following outdated defaults without understanding perceptual issues.

Key Takeaways

Colormaps convert numbers into colors to make data easier to understand visually.

Choosing the right colormap type—sequential, diverging, or qualitative—depends on the data's nature and story.

Perceptually uniform colormaps ensure color changes match how humans see differences, avoiding misleading visuals.

Misusing colormaps can confuse or mislead viewers, so understanding their purpose is essential.

Advanced users can customize colormaps for precise storytelling but must balance clarity and complexity.