6 Common Data Visualization Mistakes
Updated: May 18
“The purpose of data visualization is not pictures but insights”
- Ben Shneiderman
Data visualization uses visual encoding – by simply assigning visual attributes such as colors and shapes to different types of data. The visual encodings translates data into visual shorthand that brain can decode quite easily. These visual cues pop out to the brain with no conscious effort on our parts – making interpretation of data easier.
To communicate effectively with dashboards requires right balances in use of these pre-attentive attributes for data visualization (e.g. color hue, texture, size, shape, orientation, position & alignment, color saturation.) as well as ensuring user needs are well addressed in the visualization.
Therefore in this post I share six mistakes you should avoid while visualizing data.
1. Choosing the wrong chart/graph
Charts and graphs are the building blocks of data visualization - and your dashboards. Choosing the right chart and graph impacts immensely on how well your data story is conveyed which is true if otherwise.
This common mistake can be avoided by implementing charts/graphs which resonate with your business question. Below are some chart ideas you can implement;
When doing comparison overtime: Line chart, multi-line chart, area chart, step chart, bump chart etc.
Correlation (relationship between variables): Scatter plot, bubble chart etc.
Location analysis: Heat map, bubble map, filled map etc.
Key performance indicators (KPI’s): bullet chart, dial, thermometer, big number chart, gauge etc.
For more ideas on data visualizations, here is a free guide with 35 different ways of visualizing data to help you fast-track your data visualization skill.
2. Cropped axes
Axes provide context to our charts. Messing with the axis will completely paint a different picture than you expect. A good example is in this example below, with one axis cropped while the other starting at zero.
Looking at the cropped view, one might think phones made three times more sales than printers – which is incorrect. While the correct view (axis not cropped – axis starting at zero) gives the right perspective on the difference in sale for the two products – giving clear direction.
Such small and yet vital elements in visualizations should be evaluated on how they impact our stories before use – in cases where cropping in an avoidable, let the user know.
3. Hard to compare vizzes
The hard to compare visualizations are as a result of choosing the wrong charts and graphs for the right data. A good example is in the figure below which is comparing product sale for two different periods (Year 2018 & 2019).
Such a viz is difficult to interpret since one can’t tell the difference in slice sizes – such can be corrected by implementing the right chart for the right data. See the difference when such data is presented using a bar in bar chart.
By implementing the right chart for comparing data (2018 & 2019), interpretation is much easier and happens without much effort.
4. Displaying too much data
Nobody likes busy visualizations. So always avoid creating such. If you have too many questions to answer – always break them in to a series of visualizations instead of throwing everything in the same view.
This can also be addressed by presenting data at a course level then empowering your users with levers (filters & parameters) to slice and dice data to granular level.
5. Using metrics no one understands
Understanding user’s environment is one of the key requirements before undertaking any data visualization task. By doing so, empowers you the developer with the right vocabulary used in each specific niche – helping you develop dashboards which resonate with each specialty/department for organizations.
If not implement this could lead to visualizations which only make sense to a section of the organization – hence preventing user adoption.
6. Not viewing your dashboards as your user will
Design/develop with user in mind. Always seek to understand which device your viewer will use to interact with your visualization. Is it a desktop, tablet or mobile phone?
Understanding this will help you optimize your visualization for that device – ensuring users have same experience with your visualization across the board.
Data visualization is a vast field and if you’re reading this post it means you’re interested in getting better. Here are some resources on data visualization you can try.
I hope this article was helpful to you. To receive more of these tips on data analysis and visualization with Tableau, kindly subscribe to our mailing list.
Thanks for reading.