Visualization is not a simple data display, its real value is to design a data display that can be easily understood by readers. Every choice in the design process should ultimately be based on the reader’s experience, not the individual designer.
This article mentions some common mistakes in design, and also some tips from our team. Hopefully these 25 tips can quickly improve and solidify your data visualization design, let’s take a look!
1. Principles
1. Choose a graph that can tell a story
You first need to think clearly about what you want to achieve, what message you want to convey, and who your users are.
2. Cut out elements that are not relevant to the story
This doesn’t mean halving the amount of data, but rather beware of chart garbage, redundant information, unnecessary descriptions, shadows, decorations, etc.
The beauty of visualization is that it can reinforce and convey the story you want to convey (but don’t use 3D charts – it can skew the visual perception).
3. Design for better understanding
Once you’ve created your visual prototype, take a step back and consider how you can make the data easier for readers to understand.
What other simple elements can be added, tweaked, or removed? Perhaps adding a trend line to the linear chart, you may also find that the pie chart has too many slices (up to 6 slices).
These subtle tweaks can make a big difference.
2. Comparison
Visualization makes data comparisons more intuitive, but simply putting two sets of graphs next to each other doesn’t do the trick, and it’s even more puzzling. (Imagine comparing 32 different pie charts together? No way!)
1. Add zero reference line
Although a linear chart doesn’t necessarily start at zero, it is necessary if the chart contains a lot of comparative data!
Relatively speaking, small fluctuations in data are meaningful (such as stock market data), then you need to truncate a range to show their difference.
2. Choose the most efficient visualization
Keeping the visual consistency so readers can tell at a glance means you might want to use a stacked column, grouped bar, or line chart. But no matter which graphic you choose, don’t make the reader struggle to compare too many things.
3. Pay attention to the placement
If you use two nice stacked bar charts for the reader to compare, but if they are far apart, don’t talk about comparison.
4. Tell the full story
Maybe your fourth quarter sales were up 30%, isn’t that exciting? But there are even more exciting ones! Comparing the figures for the first quarter, sales have grown by 100%.
3. Context
True, data is about numbers, but it is often contextualized, generally to provide context for the main points that follow. But in so many data visualizations, infographics, and ebooks, we see data visualization and context in opposition, not in combination.
1. Don’t over explain
If the context already mentions something, there is no need to repeat it in the subtitle, callout, title.
2. Keep chart titles short and to the point
There is no need to use playful, long-winded or pun intended. Descriptive titles above the graph should be concise and directly related to the graph below. Remember: focus on what makes people understand quickly.
3. Make good use of callouts
Callouts are not meant to fill in the blanks, but to emphasize relevant information or provide additional background.
4. Don’t use distracting fonts or elements
Sometimes you do need to emphasize a point, just use bold or italic text, not both.
4. Color
Used properly, color is an excellent tool. But improper use can not only distract the reader, but even mislead. Therefore, use colors sensibly.
1. Use one color to represent the same type of data
If the bar chart shows monthly sales data, then just one color is sufficient. If you want to compare this year’s sales data with last year’s sales data on a set of charts, you can use different colors to represent the data for different years.
In addition, an accent color can be used to highlight important data.
2. Pay attention to the expression of positive and negative data
Do not use red for positive data or green for negative data. These color associations have historically been strong, and they have long been marked in the minds of readers.
3. Make sure there is enough contrast between colors
If the colors are too similar (such as light gray versus lighter gray), it can be difficult for people to see the difference. Instead, avoid strong contrasting colors, such as red and green or blue and yellow.
4. Avoid Patterns
Stripes and polka-dot patterns sound interesting but can be very distracting. If you want to differentiate on maps, etc., use the same color with different saturation and solid lines.
5. Use the right colors
When some colors in a chart stand out more than others, it adds unnecessary importance to the data. Therefore, a single color with distinct shades or two similar colors in the same spectrum should be used to differentiate intensities.
Remember to use intuition and color shading to adjust importance.
6. Don’t use more than 6 colors on one image
Just look at the picture:
5. Labels
Labels can become minefields. While readers rely on labels to interpret data, too much or too little can be distracting.
1. Make sure everything is labeled
Make sure all required information is labelled – and that there are no repetitions or misspellings.
2. Make sure the label is visible
All labels should be clearly visible and the corresponding data points can be easily identified.
3. Lines can be marked directly
If possible, include data labels in the data points. Readers can quickly identify lines and corresponding labels without having to look for legends or similar values.
4. Don’t over-tagging
If precision of data points is important to storytelling, include data labels to enhance comprehension. If it is not important, ignore the data label.
5. Do not place labels at an angle
If the labels on the data axis are too crowded, consider removing other labels on the axis to make the text flow more comfortable.
6. Sort
Data visualizations are designed to aid understanding, and incomprehensible random patterns are frustrating and can undermine what is meant to be conveyed.
1. Arrange data visually
Charts should have a logical structure that arranges data in alphabetical, sequential, or size categories.
2. Sorting coherently
The order of the legend should be consistent with the order in the chart.
3. Sort evenly
Use natural increments (0,5,10,15,20) on the axis instead of uneven increments (0,3,5,16,50).
