Effective Data Visualization
Data shared via text can be confusing (not to mention boring), but data presented in a visual format allows people to extract meaning from that information more quickly and easily. Data Visualization is very important because it helps us to understand the data in a better way. When we see a bunch of tables with the number, we might be able to read those tables and the number, but we can’t easily make sense of that data. Our brain understands the information better when it is shown in picture format.
Analysis has shown, a high-quality graph has a 30% more chance of being read than a text. Though there are many advantages of data visualization, at times it can be misleading and confusing too if not used properly. Data visualization is a way of storytelling. If we miss that point it will look like a bunch of shapes in different colors. From this report, we want to analyze different graphs and discuss what are the best practices to make the visualization very effective.
Example of Good Data Visualization
Visualizations which are based on time series data are generally represented in line graphs. Instead, this graph shows the average number by minute, by hour and by week. Colors used in the graphs look so appealing with the gradient. Information written in text is easily readable. Information which is very important has jet black color and bold in font whereas the information which are side notes are written in gray color.
Every figure has appropriate titles and labels along with the specific pointers. The key information about the graphs has stated at the very beginning of the visualization in a very clear and precise way. When we start looking at a graph our eyes move in uniform motion collecting and understanding information.
The data that stood out to me from fig 1 was those peaks where the value is above the average in each graph. The choice of position and shape attributes among all pre-attentive attributes is really making it eye-catching.
Along with position and shape color hue and color intensity is used very properly which depicts the important data very well. For example, if we consider the first graph Babes born by Minute our eyes get focused on the peak shown in red. Which explains that at this point the value is greater than average value. As the data in the first graph is averaged and used in minute graphs and so on, we can see the change in peeks. Patterns in the born by minute graph are skewed and patterns in the bye week graphs are smoother.
Example of Poor Data Visualization
If we can see data visualization example of MLS salaries in 2013. At first glance we see multiple color boxes with information written in very small font size making it difficult to understand. This single visualization has two subgraphs in it which don’t have any titles. We need to read the text information written on graphs to understand what these two-graphs are trying to represent. The color code information has abbreviations written making it hard to understand what word it would be. Large bar graph on the left does not have an x-label.
In terms of interpretation, it doesn’t show any pattern or trend of values in the graph. If the bar chart was sorted it would have looked more attractive and our eyes would have flown more naturally. This current chart is making our eyes roll too frequently without giving time to focus.
Bar graph on the left has multiple small boxes in a single graph, some of them are labeled and others are not, making it very difficult to understand the purpose. In this graph pre-attentive attributes like color and shape have been used but that’s what makes the graph too clumsy. Color intensity and color hue attributes could have been used to make the viewer show the important information from the graph.
One very important difference between the good and bad example of the data visualization is the data being used for the graph. Data in fig 2 looks all over the place, does not seem to be property filtered as compared to the fig1 data which is very crisp. Few points which could have been differently are instead of taking all the state information we can just show the top 10 states. Second bar chart could have been made more informative with data filtration and proper title. Key information about color code must be written in full forms.
Key Things to Remember:
- Selecting the appropriate visualization: To effectively communicate the information, selecting an appropriate visualization type, which aligns with the data and analysis is necessary.
- Keep it simple: Avoid cluttering the visualization with too much information or using overly complex designs.
- Use colors effectively: Colors can be used to highlight important information or to create a sense of hierarchy, but they should be used judiciously.
- Provide context: Make sure the visualization includes enough information to allow the viewer to understand the data, such as the time period being shown or the units of measurement.
- Consider the audience: Think about the people who will be viewing the visualization and tailor it to their needs and level of understanding.
- Validate your data: Make sure the data being used is accurate and reliable before creating the visualization.
- Be consistent: Be consistent in your use of colors, labels, and other design elements throughout the visualization.
- Annotate: Add labels and annotations to help explain the important features of the visualization
- Tell a story: Use the data to tell a story, and make it easy to understand and interpret.
- Iterate and test: Always test the visualization with different people and gather their feedback, and iterate on the design until it effectively communicates the data.
Before moving to start visualizing the data, its helpful to pause before creating a dashboard and reflect on what message we want to convey and what elements are needed for it. These two graphs discussed above show that distinguishing between what you really need and what you don’t reduces the data visualization as much as possible so that it contains only the main parameters and provides both information and context.
References
1. Nadieh, B., Zan, A., Jennifer, C. (2017). Why Are so Many Babies Born around 8:00 A.M.?. Retrieved from https://www.informationisbeautifulawards.com/showcase/2100
2. MLS salaries in 2013. Retrieved from https://public.tableau.com/views/MLSSalaries/MLSPUDashboard?%3Aembed=y&%3AshowVizHome=no&%3Adisplay_count=y&%3Adisplay_static_image=y
3. Why Visual Analytics?. Tableau Blueprint Help. Retrieved from https://help.tableau.com/current/blueprint/en-us/bp_why_visual_analytics.htm