Evolytics Blog

Great data viz is tool agnostic

Great Data Viz by Evolytics

A process for creating successful dashboards and visualizations in any tool

Data visualization is often considered synonymous with a report, a graph or a specific tool, such as Tableau. However, data visualization is more than a tool, chart or output. Rather, it is a process of thinking about how best to quickly and accurately convey important information. This goal can be accomplished multiple ways, through a variety of tools, and ultimately requires thinking through a set of questions and objectives. The decisions involved to create a successful data visualization fall into three categories:

  • Organization
  • Design
  • Interactivity

These areas must be considered from the beginning of a project – acquiring a data set – to the end – the final visualization product. The key to getting value from data visualization is not the selected tool but knowing which data visualization decisions to make. The first, and often overlooked, decisions to make involve organizing the information at hand. This process should happen before ever opening the visualization tool because it defines what success looks like. The saying “if I had more time, I would have written a shorter letter” also applies to data visualization. It is easy to throw every single data point onto a dashboard and call it done, but it takes time and thoughtfulness to narrow down which metrics and dimensions are most important. To successfully complete the “organization’” stage of creating a data visualization, try practicing these steps using the example of an online retailer:

Who is the audience of the dashboard?

Answering this question allows one to figure out the total number and types of stakeholders. If there are multiple stakeholders, it may be more appropriate to create two separate dashboards.

Example Stakeholders: 

  • CEO
  • Website analyst

What question(s) is the visualization or dashboard trying to answer?

Knowing what the audience is trying to figure out from the dashboard helps narrow down what information to display and ensures that answers are easy to find.

Example Business Questions:

  • CEO: How is our website performing generally? Are people actually purchasing our product?
  • Website analyst: How are people getting to our website? Is paid or unpaid marketing more successful in getting people to purchase? What are people doing on the website before they purchase? Are people able to find the products they are looking for?

What metrics and dimensions are available?

Making a list of every single metric and dimension that is, or could be, relevant helps get a lay of the land. This variable list makes it easier to recognize the amount of information available and figure out which data points make sense to include or not include in the dashboard.

Examples:

Dimensions Metrics
  • Marketing channel
  • Device type
  • Page viewed
  • Date (month, week, day)
  • Location
  • New or repeat visitor
  • Unique visitors
  • Purchases
  • Account signups
  • Number of internal searches
  • Page views

What metrics make sense to group together?

Dividing metrics into separate categories helps define what information belongs together on the dashboard. Example categories might include “key performance indicators (KPIs)” and “secondary metrics.” Or categories could split the subject matter logically, based on business objectives such as acquisition, conversion, and retention.

Example Metric Categories:

Main KPIs Secondary KPIs
  • Unique Visitors
  • Purchases
  • Account Signups
  • Number of internal searches

How should I break out my metrics by different dimensions?

Aggregate numbers are helpful to define what happened. However, stakeholders will likely want to know what a number represents granularly. Multiple dimensions could potentially provide explanations for a number. In the list of data points, note which metric and dimension combinations would provide value to the stakeholder.

Example Metric Breakouts:

  • Unique Visitors – date, marketing channel, device type
  • Purchases – date, marketing channel, device type
  • Account Signups – date, marketing channel
  • Number of internal searches – date, marketing channel, device type, new vs. repeat visitor

How should space be maximized?

Dashboard real estate is valuable! Stakeholders don’t want to scroll through lots of duplicative information to find what they need. You can maximize space multiple ways, including filters, drop-downs for choosing metrics or dimensions, or consolidating related data points that answer a specific business question. Example ways to maximize space include:

  • Create one chart for channel, one for device type, and a trend line for date. Allow the user to choose a metric from a dropdown menu to populate graphs. For example, choosing “visitors” would show visitors by channel, new visitors vs. repeat visitors, and visitors by date.
  • On a trend line, allow the user to switch between date granularities of day, week or month.
  • New and repeat visitors do not need to be visualized separately, but can instead be included as a dimension filter to apply to dashboard charts.

How should data points and charts be organized on the dashboard?

Example Dashboard Mockup

Example Dashboard Mockup

The last phase is to mock up the general layout. If a filter is applied to the entire dashboard, it often helps to place it along the top or left side. Call-out numbers are also easy to find when placed towards the top. This allows stakeholders to see the most important information first and then have the ability to dive deeper. Creating a mockup, wireframe, or sketch allows you to quickly move things around and figure out what works and what doesn’t… before ever jumping into a data visualization tool!

Organizing the layout of data points is critical to creating a successful data visualization. Thinking through these steps will help build a solid foundation to create the end product in your preferred data visualization tool. Stay tuned for the next post in this series about the different design decisions necessary to create successful data visualizations!

Lindsey Poulter

Lindsey Poulter is a data visualization analyst specializing in Tableau. She helps clients gain insight into their performance through visually appealing, interactive dashboards. Lindsey created a Top 5 Tableau Public Visualization of 2016 and is a certified Adobe Analytics Business Practitioner.

Lindsey PoulterGreat data viz is tool agnostic

Related Posts