Below we discuss a number of different encoding features and how to determine which to use when designing charts. Encodings are the visual cues we use to represent data on a chart. They can be shape, size, color hue, color intensity, position, angle or orientation, icons, isotype, and text. Choosing the right encoding is essential for creating accurate and easy-to-understand visualizations.
Shape: Consider using shape encoding when you need to distinguish between different categories or groups within your data. For example, when comparing the sales of different products in a scatter plot, you can use different shapes, such as circles, squares, and triangles, for each product category. This will help viewers quickly identify the different products on the chart.
Size: Size encoding works best for representing continuous variables, such as population or revenue. For example, in a bubble chart showing the populations of various countries, larger bubbles can represent larger populations. Similarly, taller bars would correspond to higher revenues in a bar chart depicting company revenues. Remember, using size can be tricky as humans tend to perceive differences in area rather than radius or length.
Color hue: Color hue is another way to represent categorical data. You can use different colors to represent different categories or groups of data. For instance, in a pie chart that represents market share, each segment can be a different color to represent a different company. Be mindful of color blindness, and choose distinguishable colors that don’t confuse the viewers.
Color intensity: Color intensity can be used to represent continuous variables when you have a single color and want to show changes in intensity or value. For example, in a heatmap showing temperature variations, you could use a gradient from light blue (low temperature) to dark blue (high temperature). This encoding helps to show the relative difference between data points.
Position: Position is a powerful way to represent continuous variables. For example, bar charts utilize the position of bars along the vertical or horizontal axis to represent values. Line charts use position to represent changes in a variable over time. The position of data points on an axis determines their value, making it easy for viewers to grasp quantitative differences between data points.
Angle or orientation: Angle encoding can be useful for showing proportions in a pie chart, where the angle represents the percentage of each category. Similarly, on a slope graph, the orientation of the lines connecting the two points gives a clear view of whether the value is increasing or decreasing.
Icons and isotype: Icons can help represent qualitative data, while isotype charts use repeating icons to indicate the quantitative aspect of data. For example, you can use different icons for different transportation modes in a chart comparing the number of users of different transportation systems.
Text: The last but equally important encoding is the use of text. Labels, legends, and annotations provide additional context to your chart, helping viewers understand the data better. For example, adding axis labels and a chart title can significantly improve the overall interpretability of a bar chart.