Data Visualization

This guide collects resources and tutorials related to data visualization.

Best Practices for overall great designs

  • Design your visualization for a specific audience and identify their needs.
  • Make sure your data is clean.
  • Provide a context for your visualization. This can be done with good titles and labeling
  • Choose the right visualization for your needs.
  • For most purposes, keep your visualizations and dashboards simple and easy to understand quickly.
  • Format your visualization for accessibility.
  • Highlight the important parts.

Avoid these perils

There are a number of practices that are perilous. Some make your visualizations harder to understand and some may misrepresent your data.

Do not use partial axes. For bar charts and many line charts the numerical axis should start at zero. Readers will draw inaccurate conclusions when looking at truncated charts.

Do not skip numerical values when you are using a full axis. Use consistent intervals.

Do not confuse the viewer with extraneous symbols. Gridlines, axis labels and colors can all be simplified to highlight the most important information.

Do not just accept the out of the box methods for labeling. Put labels near the lines of your line graphs. Rotate labels and bars to make reading more comfortable. Simplify axis labels if possible (for example using mil. for millions in lieu of all of those zeros).

Do not make your users struggle - they should not have to squint to see the details. If you are projecting your visualization remember that projectors will wash out your colors - use high contrast designs.

Do not just rely on your eyes - have others look at your draft visualizations.

Reconsider your use of 3D or blowout effects. Studies have shown that these techniques inhibit understanding.

Do not go crazy with color. Color is a valuable tool but limit the number of colors you are using to no more than six. Consider using color families to identify categories. Some colors are difficult for persons with color blindness to perceive. Check the readability of your visualizations by printing them in gray scale which helps to determine the hue and saturation of your colors.

Do not change style in midstream. If you are creating panel charts or using charts for comparison, it is a good practice to use the same format, colors, axes, and labels.

Do not cram too much into your visualization. Try changing chart types, removing or splitting up data points, simplifying colors or positions.