Best Practices for Data Visualization
Even before the written script was invented, humans used to communicate through paintings. We understand better and absorb more when the information is presented in the form of visual elements. In simple terms, data visualization is representing the data in the form of various visual elements like charts, graphs, etc.
This graphical representation of data using data visualization has many advantages. Looking at data visualization makes it easier to see and understand the patterns, trends, outliers and relationships in the data even for non-technical audience. Data visualization can be made interactive. It makes it very easy to share information.
In this article we will look into the best practices for Data Visualization to improve user experience.
Decide the story
The first step towards creating a great visualization is to decide on the story. Clarity on what we want to convey to audience is very important before starting the data visualization project. All the key insights of the data that we want to convey should be already identified and the visualization should be used to highlight them in a way that it is easy for the audience to understand and make inferences from the data visualization.
Choose right chart
Once the story is fixed, we have a fair idea of what we want to convey to the audience. Now the next question is — how. Therefore, choosing the right chart type is very important in conveying the information in any data visualization. Each chart type is suited for a specific type of data and use cases.
Bar Charts, for example, are used to make the comparison of data across multiple categories. Visually it is very easy to notice and make inferences related to difference in data using a bar chart. Line charts, on the other hand, are used to show trends over time. Maps are used to show demographic data.
Why are Chart Type important?
Let me explain using an example. Suppose we have data related to a specific category of items, say washing powder, sold across stores of a retail chain. The category has multiple types of washing powder like soft, scented, normal, etc. each representing a specific SKU.
If we want to compare the sales of different SKUs in the category for any given month, we will use Bar Chart. If we want to analyze the sales of a SKU across a few months or an year, we will use line chart. If we want to understand the which type of washing powder is used in north of country and which is used in south of country we will use map to denote this information.
Make everything as simple as possible
Albert Einstein has said “Make everything as simple as possible, but not simpler”. While creating data visualizations, it is important to keep things simple. Adding too many colors, labels, or other elements can make the visualization hard to read. A simple and clean design will ensure that the audience will understand the key insights of the data and take away correct message from visualization.
Color is a powerful tool to draw attention to specific information in any data visualization but it is also important to use color consistently and in a way that supports the message we are trying to convey. Using a consistent color palette and font throughout the visualization will ensure that the audience is not overwhelmed by colors.
Limiting the number of elements used in each chart will ensure the clean chart thereby the audience can focus on the storyline.
Always label Data
Labels and legends provide the context and make it easy for the audience to understand that the data is representing and drive the message straight to the audience. As a best practice, always provide labels for x and y-axis and add labels for individual data points (in case this does not clutter the visualization too much).
Use Dashboard to drive message
Dashboard is the place where we stich different charts, graphs and other visualization elements to tell the story. Dashboard is usually used to provide overview of specific topic and acts as a placeholder for providing context to the data visualization and allows users to explore the data in more depth. Dashboard also acts as the first source to monitor KPIs and make initial decisions based on the data.
Last but not the least — ITERATIVE TESTING
It is always good to test the data visualization with a small group of users and get feedback. This process helps in identifying issues and improvement areas that can be implemented to create a more effective visualization. Getting the feedback from a small group of users shortens the feedback loop making the improvement of data visualization faster.
Conclusion
The data visualization project must start with choosing a storyline. This helps in choosing correct chart type. Consistent color palette and font are very important aspects related to look and feel of the visualization. Keeping things simple with clean chart is also important. We should always provide context with data labels. And then finally collate the related charts and graphs in a dashboard which can also act as overview of the information. And finally, always test and iterate with small group before publishing the data visualization to a larger audience.
