4 Best practices for designing data visualizations
Introduction
During my Spring 2020 internship at PointClickCare, I applied our visualization library patterns to visualizing vitals for nurses that work in nursing facilities. I learned to appreciate the power of data visualization, and understand best practices for visualizing data. Here I share 3 advantages of data visualization and 4 best practices I learned which are simple concepts but powerful guidelines to follow.
3 Advantages of data visualization
1. Surfacing critical insights.
Even though people like data, but people do not like staring at numbers, especially a lot of them at once.
2. Making interpreting large volumes of data quickly.
Line charts for example, removes the unnecessary details and weave connections among all the numbers in a table to make it simple for a user to analyze.
3. Presenting patterns and pattern violation.
Our brain is skilled at detecting patterns and violations. We can easily discover trends, gaps and outliers that are presented graphically.
4 Best practices for designing data visualization
The following are 4 guidelines I learned from my manager to visualize data.
1. Avoid “circular” shape graphs for part to whole comparison.
Visualization experts recommend to avoid circular graphs such as radar graphs, circular graphs, pie charts, and gauges etc, because our brain has difficulty distinguishing angles and arc-lengths, in comparison to straight line charts.
2. Use colors intentionally.
There are 5 cases of how to apply color:
a) Sequential colors
Increase in color intensity conveys a magnitude from low to high.
b) Diverging colors
Two sequential colors with a midpoint to convey a spectrum, such as cold to hot.
c) Categorical colors
Contrast colors for individual comparisons. Colors have to be used in a defined order.
d) Highlight colors
One soft color to highlight something.
e) Alerting colors
One intense color (often red) to warn or alert the user.
3. X/Y axis title labels are optional.
Power users can find title labels redundant and unnecessary; or in cases where graph titles or other context make the labels obvious and there are too many information on the page, it is better to hide the title labels.
4. Data should take up no more than 80% of the graph’s Y axis.
Whether that’s scatter plots, line charts or bar graphs, the most effective use of vertical space, and most appealing to the users is 80% of Y axis. Ideally the graphs would adjust according to the data being inputted for the users.
Conclusion
By working in a field where data couldn’t be leveraged because it was poorly documented in the past, it made me retrospect on the way of how we are storing our own “data” and leveraging our “data”, such as our notebooks, journals, notes from school etc.
Are you documenting or generating “clean data” to leverage and take advantage for yourself in the future one day?
