How To Build Data Visualizations with Your Goals in Mind

There are millions of ways to format and style a data visualization, so how do you build yours to be the most effective?

Visualization purists, such as Edward Tufte, advocate for minimalistic, data-heavy graphics which are stripped of any gratuitous elements. While this is often the case, we should keep in mind what and to whom we are presenting – there are times when finding patterns within the data is not the main goal.

Currently at Umbel, we have two major ongoing projects – refreshing the brand style and redesigning our product. In our efforts to maintain the cohesion between data visualization in both in our brand, and data visualization in our product, we often find ourselves asking how the goal defines what format will be the most effective.

For example, we recently hosted a SXSW Party called UMBELMANIA featuring a full day of panels and Spoon at ACL Live. We had over 10,000 attendees register for the event. Let’s walk through how we might represent the RSVPs to our UMBELMANIA party over time – in both a Marketing and a Product context.

In the product, the goals of visualizing RSVPs over time might be:

  • See when (day of week, time of day) people were most likely to RSVP
  • Learn what promotion tactics did and did not work
  • Illustrate patterns in the data that help Marketers update their strategy

In order to meet these goals, we want to build a clean, minimal data visualization that focuses on giving the user control. The user should have the ability to control the sampling interval, in order to find daily, weekly, and monthly patterns. They could also have the ability to brush, or zoom in, to get more detail about a specific day. Displaying detailed information on hover could give the user more information about each data point, but only when they need it.

Here is an example of how we visualize this information in the product:

In Marketing materials, the goals of visualizing RSVPs over time might be:

  • Build brand awareness
  • Get people interested in your content
  • Illustrate patterns in the data

In order to meet these goals, we want to build a very different data visualization that is visually appealing and will draw users in. However, keep in mind that the shift in focus from the data to the presentation does not give an excuse for misleading or obfuscating the data. We could use the brand colors, motion, and the floorplan of the theater to create a more vibrant, exciting graphic.

Here is an example of how we might visualize this information:

Here, instead of showing numbers or the rate of RSVPs over time, we have a sense of the cumulative number of RSVPs in respect to the venue itself. To read a number in this visualization, the user would have to count dots — a completely inefficient way to display! However, you can see that we filled the venue within two weeks and we had more than 3x the number of RSVPs as capacity of the venue. We are giving the viewer a sense of the relative scale, as opposed to a clean view of trends over time.

While there are core principles for data visualization, there are multiple good solutions for visualizing the same data. So remember — defining the end goals, and keeping those in mind, is a key factor in an effective graphic.