In our last post, Dashboards 101, our intrepid analyst Regina was building a dashboard for her company’s management team. Regina works at a fictitious coffee delivery company named Mocha. One of Mocha’s projects is to acquire more specialty coffee houses to their client roster. They would like to cater to customers who shop at organic, paleo friendly, and vegan coffee shops.
Regina is putting together a dashboard for management to review. She scheduled a meeting with a couple of sales and marketing managers to get their take on it before presenting to senior management. Here's the first draft:
What is data visualization?
Quick refresher: a dashboard is defined as a type of graphical user interface which often provides at-a-glance views of key performance indicators relevant to a particular objective or business process.
The visual elements on the dashboard such as the charts and graphs are data and information being represented as a graphic. Data visualization is the graphical representation of data. It allows us to see and understand trends and patterns in our data.
In the end, data visualization is a mode of communication based on visual representation of data. The goal is to make it easier to convey information—and insights!—to a variety of stakeholders.
What data visualization is not
In the world of data, terminology can easily become interchangeable erroneously. Data visualization is often confused or mistakenly replaced with the following terms:
Business intelligence - I like to view business intelligence (BI) as the ecosystem or analytics stack. It’s the platform where analytics, data mining, data visualization and other data tools come together to help organizations make data-driven decisions. So data visualization is the graphical component of BI.
Data Analytics - I like to compare data analytics to data sleuthing. Analyzing data allows us to search for trends and patterns in order to draw conclusions. We use data visualization to present those findings in a way that makes the results easier for broader audiences to understand.
Data Literacy - Data literacy is the ability to read, understand and communicate data. It is often confused with data visualization more often than the other terms. Being familiar with data is different from being skilled at visualizing it.
Gestalt principles and data visualization: TL;DR, less is more
It’s been my experience with data visualization that less is more. For broad audiences, I find simple charts in a decluttered space are received well. For the most part we intuitively know what seems right when looking to communicate a point or idea visually. Generally speaking, we all crave some form of structure.
There is a branch of human psychology that has been studied for a very long time called gestalt principles. Gestalt is a German word that translates to “unified whole” in English. It’s based on the idea that the whole is greater than the sum of its parts.
Not only do these principles validate our need for structure, they help us improve our data visualization techniques. I'll walk you through them and provide examples—take a look:
The law of simplicity
Simple charts are better for communicating data. They are easier for audiences to process and remember than visually complex charts. When creating visualizations, bar and line charts are recommended.
This chart is simple, but that's the point: By using a familiar chart type and using color coding, it's easy for viewers to stay focused on the most relevant information.
The above chart shows monthly sales transactions for the first three months of 2021. Using techniques like sorting items, removing the use of multi-color, and simplifying or reducing jargon helps the audience process what they are viewing much easier. The above chart is clean and emphasizes the most recent month on the chart. This provides focus on what is relevant.
The law of proximity
Positioning is very important when visualizing data. It helps to convey relationships between data points. When objects are close to each other they are seen as part of a group versus when they are spaced farther apart. We can present information by organizing the space around each group.
Because the data for specialty coffee houses ends up grouped, it's easy to see that non-specialty revenue rises above the rest. But just as important, the legend placement makes deciphering the chart easy.
The chart above shows revenue trends for Mocha clients by specialty. The legend across the top shows us the specialty groupings that Mocha is looking to grow. Thanks to proximity, the takeaway that specialty coffee houses lag behind general or non-specialty coffee houses is easily recognized
Another example of leveraging positioning is the placement of the legend across the top versus placing it to the right. This allows for the audience's eyes to not have to make as much movement to identify which label corresponds to the line. Another recommendation is to place the label directly on the line.
The law of focal point
The focal point principle states that whatever stands out visually will be given priority attention by the audience. One successful way of achieving this is to use color to our advantage.
The chart to the left shows the sales revenue for first quarter 2021 for select specialty coffeehouses. During Regina’s meeting with Sales, they were particularly interested in comparing the revenue of Dave’s Organic Beans and A&E Coffeehouse.
Using color, Regina was able to put the focus on the two coffee houses in question while still providing crucial context about other locations that the Sales team needs to keep tabs on.
The law of figure and ground
The law of figure and ground states that we tend to segment our visual world into a point of focus and background.
The figure and ground principle describes the capacity to perceive the relationship between form and surrounding space to create meaning. Put another way, a sense of wholeness or unity depends on how you perceive the relationship between an object and the area in which it is contained.
Adding shaded reference bands to the background of these charts makes it easier for Regina to provide crucial context for understanding the data without distracting from key takeaways.
In the above chart Regina starts to explore if there is any relationship between monthly sales revenue and delivery time and distance for select specialty coffee house clients. To help with the analysis, a reference line was used to mark the average revenue at $60K and reference bands were used to give the minimum to average range for both delivery distance and delivery time.
By using a faded color for the average bands, it fades into the background so viewers can focus on the data points in the foreground. As a result, it’s easy to see that Dave’s Organic Beans has high revenue numbers and a minimum delivery distance compared to the other customers.
The law of similarity
This law states that we seek similarities and differences and tend to categorize similar items in groups. Similar elements are visually grouped, regardless of their proximity to each other.
We can use this in data viz by relying on similar characteristics (color, size, shape, etc.) to establish relationships between objects and to reinforce groupings.
If you find this waterfall chart difficult to read, it's no wonder: Regina hasn't properly applied the rule of similarity.
In this example, Regina is looking to analyze the number of net adds of specialty subscribers to the Mocha platform. In this first draft, she grouped subscriber additions using shades of blue and green to organize the specialty categories. Meanwhile, churn is coded in red and yellow.
Although this waterfall chart appeals to the law of similarity, it’s still hard to digest. There’s a lot going on, and Regina’s colleagues felt the use of color and having to read the legend was confusing.
With that feedback in mind, Regina decided to pair the law of similarity with the law of simplicity. She tried organizing the data in a bar chart to make it easier for the audience to understand subscriber adds and churn.
This is the same data as in the chart above, but it's a lot easier to read now that Regina has applied both the law of similarity and the law of simplicity.
While this chart is based on the same data as the waterfall chart above, it’s a lot simpler. With fewer colors and intuitive above-and-below bars, it’s easier for Regina’s colleagues to see that pet-friendly and vegan subscribers are leading the pack, even taking churn into account.
Creating a cohesive dashboard
Now that Regina has her charts in order, she’s ready to pull them into a dashboard that Sales and Marketing can use to guide their decisions.
Dashboard consultant Bridget Cogley likes to draw on her experience as an American Sign Language interpreter when she advises on dashboard design. When it comes to designing dashboards, she advises that accuracy is more than a chart level decision.
She advises, “It’s the charts, the way we link them, and everything that ties it together that comprise a visualization or dashboard.” When creating a dashboard, you want to make sure that it’s cohesive and the same gestalt principles that apply to individual charts can be used to ensure that everything works together well.
Regina’s meeting with the Sales and Marketing managers turned out to be very productive. She tested her dashboard with stakeholders and got more information about what the business teams were looking for.
By prototyping her charts and getting feedback from real users, Regina was able to create a much more cohesive—and useful!—dashboard.
The above snapshot is a reflection of those meetings. She learned that the team is very focused on San Francisco neighborhoods at the moment. As they bring on more coffeehouses to the Mocha platform, they want to track delivery distances and times to see if there is any correlation between these measures and revenue.
Because Regina took the time to present data in a way her teammates could understand, they now have an at-a-glance guide for targeting future Mocha customers that also provides insight to their current ones.
Putting it all together
As we follow Regina’s journey, we can see that she hasn’t built her dashboards overnight. Here are some tips for better data visualization:
- Know your audience. I can’t emphasize enough the importance of working with your audience to understand what they need and are able to comprehend.
- Flesh out your story. Spending time on identifying your audience’s needs and creating a story arc will make your visualization journey much easier. Don’t get too bogged down with what data visualization tools you’re going to use in the beginning. When it’s time to use your data visualization platform of choice, come with your story.
- Know your data. Now that your story has been fleshed out it’s time to map out what data is needed and make sure it’s available and accessible. The more granular your data, the more flexible it will be for building out your visualization.
- Less is more when visualizing. Make sure to choose the right chart for your analysis. In addition, be mindful of other principles that will help your audience better navigate your visualization like proximity, grouping, focus, and color.
- Design in grayscale. My mind was blown when Bridget Cogley shared that she designs her dashboards in monochrome and then considers color. This approach is great for focusing on the shape of the data and forces us away from complexity and toward other techniques such as interactive visualization.
- Make it consumable. When designing your visualizations, keep in mind how they will look on various screens. Does your audience do most of their viewing on desktops, laptops or mobile devices? Executives tend to review things on their mobile devices so make sure that your visualizations render well across various screen sizes.
- Make it actionable. When presenting your findings, make them simple for your audience to comprehend. While it’s your goal to highlight trends in the data, you should include takeaways and recommendations on what course of action your stakeholders should take.
Data visualization is very important in helping us explore trends in data. While visualization matters, it hinges on having accurate, up-to-date data at your fingertips. That’s where Panoply comes in!
Panoply is committed to making it easy for companies to sync, store, and access all their data. They offer code-free data connections that can be set up in minutes and the ease of a managed warehouse. Panoply recently announced that they are now available on Google Cloud Big Query. This means Panoply customers get the power of Google without bringing on a dedicated DBA, data engineer, or IT lead to own their data warehouse—get a demo to learn more!
Allen would like to acknowledge Bridget Cogley for her expert assistance with this article.
