登录查看更多内容

What are the most effective color palettes for Machine Learning data visualization?

由人工智能和领英社区提供技术支持

Data visualization is a crucial skill for Machine Learning practitioners, as it helps them explore, analyze, and communicate their data and results. However, choosing the right color palettes for different types of plots can be challenging and affect the readability, accuracy, and aesthetics of the visualizations. In this article, you will learn some basic principles and tips for selecting and applying effective color palettes for Machine Learning data visualization.

此文章中的业界达人

由社区从 3 条内容中精选。了解更多

Jean Murray de Villiers

Head of Analytics | Data Scientist | SAS Certified Across Entire Analytics Lifecycle | Helping Customers Realize…
Sergii Tkachov

Senior Staff Engineer Disruptive AMS Tools and Methodologies at Infineon Technologies
Jason Soo Hoo

AI/ML Engineering Leadership | AI/ML, Data Platforms, Software Engineering, Organizational Leadership

1 Color theory basics

Before diving into specific color palettes, it is important to understand some basic concepts of color theory, such as hue, saturation, value, contrast, and harmony. Hue is the name of the color, such as red, green, or blue. Saturation is the intensity or purity of the color, ranging from dull to vivid. Value is the brightness or darkness of the color, from black to white. Contrast is the difference between two colors, which can create visual interest or confusion. Harmony is the balance and coherence of colors, which can create a pleasing or discordant effect.

添加您的观点

Jean Murray de Villiers

Head of Analytics | Data Scientist | SAS Certified Across Entire Analytics Lifecycle | Helping Customers Realize Business Value with AI & Analytics.
举报内容
I follow these three guidelines when selecting colors for a dashboard: Color Harmony: Choose color combinations that are harmonious and visually appealing. Complementary colors (opposite on the color wheel) can be effective for highlighting contrasts, while analogous colors (adjacent on the color wheel) provide a more harmonious feel. Contrast: Use high contrast between data elements and the background to enhance visibility. High contrast can help distinguish data points and make text or labels legible. Avoid Overloading: Limit the number of distinct colors in your palette to prevent visual overload. Too many colors can make the visualization confusing. Instead, use color to group or differentiate data categories logically.

已翻译

赞

2 Color types and roles

Depending on the type and purpose of the data visualization, you may need to use different types and roles of colors. For example, categorical colors are used to distinguish discrete groups or categories of data, such as countries, products, or classes. Sequential colors are used to show ordered or continuous data, such as time, temperature, or density. Diverging colors are used to show data that deviates from a central point or value, such as positive and negative values, or differences from a baseline. In addition, colors can have different roles in a visualization, such as highlighting, encoding, separating, or grouping data.

添加您的观点

3 Color palettes examples

When creating data visualizations, there are many sources and tools that provide ready-made color palettes, such as ColorBrewer, Matplotlib, Seaborn, or Plotly. However, it's important to check and adjust the color palettes according to your data and context. For categorical data, qualitative or discrete color palettes with distinct and contrasting hues should be used. It's best to avoid too many colors or similar colors that can create confusion or clutter. For sequential data, sequential or continuous color palettes with varying values or saturations of a single hue or a smooth transition between two hues are recommended. Rainbow or spectral color palettes should be avoided as they can distort the perception of the data or introduce artificial boundaries or gradients. Lastly, for diverging data, diverging or bipolar color palettes with contrasting hues at both ends and a neutral hue in the middle are best. It's important to avoid using diverging color palettes for data that does not have a meaningful center or zero point.

添加您的观点

4 Color accessibility and ethics

Another important aspect of choosing color palettes for data visualization is to consider the accessibility and ethics of the colors. This means that you should ensure that your colors are visible and understandable for people with different types of color vision deficiencies, such as color blindness or low vision. You can use tools such as Color Oracle or WebAIM to test and adjust your colors for accessibility. Moreover, you should avoid using colors that may have negative or misleading associations or implications for your data or audience, such as red for danger, green for good, or blue for boys.

添加您的观点

Jason Soo Hoo

AI/ML Engineering Leadership | AI/ML, Data Platforms, Software Engineering, Organizational Leadership
举报内容
Having strong contrast between categories is important for users with visual impairments to be able to understand the data. For example, it is not advisable to use a red, and slightly lighter red, to label two different categories. A red and a black would be better instead. Also, consider that some users are color blind, so color should not be the only way to identify a category. A color should be accompanied by a clear text label to make it clear to the reader.

已翻译

赞

5 Color customization and coding

Finally, you may want to customize and code your colors for data visualization to suit your preferences, needs, or standards. For example, you may want to use specific colors that match your brand identity, theme, or message. You may also want to use consistent colors across different plots or reports to maintain coherence and clarity. To customize and code your colors, you can use various formats and systems, such as RGB, HEX, HSL, or CMYK. You can also use tools such as Color Picker or Color Converter to generate and convert your colors. To apply your colors to your data visualization, you can use various libraries and functions, such as matplotlib.colors or seaborn.color_palette in Python.

添加您的观点

6 Here’s what else to consider

This is a space to share examples, stories, or insights that don’t fit into any of the previous sections. What else would you like to add?

添加您的观点

Sergii Tkachov

Senior Staff Engineer Disruptive AMS Tools and Methodologies at Infineon Technologies
举报内容
Do NOT rely solely on the colors. Redundancy is your friend. Never rely on one channel to communicate the information. Same is also here: do not rely solely on the colors to do the job. Try to think and find alternative ways to deliver the message. The good trick: imagine (or actually make) your data visualization in black and white. Can you still distinguish different clusters? Is it possible to separate overlapping lines on a graph? It is always helpful to augment the meaning of the color in some way, for example with shape. Multiple histogram bars can have different filling patterns; or various groups of point in scatter plot can have unique appearance like dot ??, square ?, star ??, etc. Colors are powerful tool but not sufficient.

已翻译

赞

Machine Learning

+ 关注

给文章评分

我们借助人工智能创建了此文章。您认为这篇文章怎么样？

很棒不太好

举报此文章

查看全部

What are the most effective color palettes for Machine Learning data visualization?

1

2

3

4

5

6

1 Color theory basics

2 Color types and roles

3 Color palettes examples

4 Color accessibility and ethics

5 Color customization and coding

6 Here’s what else to consider

Machine Learning

给文章评分

感谢您的反馈

更多Machine Learning相关文章

更多相关阅读内容

What are the most effective color palettes for Machine Learning data visualization?

1

2

3

4

5

6

1 Color theory basics

2 Color types and roles

3 Color palettes examples

4 Color accessibility and ethics

5 Color customization and coding

6 Here’s what else to consider

Machine Learning

给文章评分

感谢您的反馈

查看其他技能