Humanizing data visualization

August 27, 2018

Tags: Oracle, Aws, Technical Track, Cloud, Rds, Hugepages

What is the ultimate goal of data and why is it so important? From the dawn of civilization, data has been the key to power. Ancient Indians captured data from living experiences and captured them in the Vedas and the Upanishads, giving birth to one of the most advanced civilizations thousands of years before the Egyptian pyramids. Alexander the Great built the Library of Alexandria to house ‘all the knowledge in the world’ under one roof. Subsequent emperors made it a rule that all ships must turn in their scrolls for copying in exchange for permission to dock in Alexandria and trade with their wealthy population. Today, data is causing all sorts of disruptions and political upsets across the planet. But not all data is created equal. Knowledge is power. The quality of knowledge, one that will give you the right kind of powers, ultimately depends on the data you have available. Indeed, lots of data is useful. However, when it comes to user interfaces, too much data can take away users ability to make the right decisions. A tsunami of data is allowing AI and algorithms to transform the world we live in. But humans can only handle a limited cognitive load. So interfaces that are designed for humans must be precise and empowering. Interfaces for data visualization must enable the user to focus on the right set of data to make the decisions that will produce insights and the desired outcomes. The high availability of data and sexy visualizations make it tempting to overburden users with slick looking, data-packed dashboards or overpopulated tables that are ultimately meaningless. Users typically approach data, reports and dashboards with a specific goal in mind. The ideal user experience makes the interface invisible, offering the user the necessary data at a precise time to complete a goal or an action. For example, imagine an enterprise user whose role is to manage data and reduce IT risk . While they might need access to vast amounts of data generated by their IT infrastructure, what is truly valuable is the ability to respond to threats or events using real-time notifications and altering trends. Remember, we’re talking about designing interfaces for humans here. Think about what your typical workday looks like: the interruptions, meetings, breaks and other events. You need to be able to pick up where you left off. User interfaces that house the data should aid your work rather than become an additional roadblock. Let me be clear, I am in favor of data visualization, dashboards, tables, etc. As the Head of Product Design for Tehama , I am always advocating that we design for humans. Here’s a seven-step approach you can take to create meaningful data visualization experiences .

Step 1 - get to know users

Create personas for the user groups that will be using the data you want to deliver. Through observing or creating an empathy map, understand how the user may think or feel before, during and after they interact with your system. What is their role? The use case? What are their pain points and key objectives? Learn as much as possible about each persona.

Step 2 - design a user journey

Using everything you have learned about users, walk in the shoes of each persona by noting the steps a user might take to complete key tasks. Don’t spend time thinking about the UI here. Simply map the steps. Once you have a basic journey identified, review it and see what’s missing or what is unnecessary. One way to refine your user journey is to create an emotions map in parallel with your user journey. Put yourself in the shoes of the user and imagine how the user would feel at each step. You can use happy, indifferent and sad face emoticons for representation. Try and increase the number of happy faces and reduce sad faces to as few as possible as you iterate on your user journey.

Step 3 - analyse the value chain

Once you have identified the type of data the user might want to see on their screen, ask yourself, “why?” or the ways in which the information will assist the user. Determine the value it creates and then work backward to determine how the user interacts with other users or individuals in their organization. An analysis of the value chain offers ways to create value for users along the value chain and uncovers important clues about relationships between user capabilities and intentions.

Step 4 - create quick prototypes

Transform abstract ideas into tangible artifacts. This will help you test your assumptions and share your ideas and gather feedback from stakeholders and potential users. Rapid prototyping techniques include storyboarding, user scenarios and experience journeys. The cost of a simple 2D prototype can be as low as a pen and some paper. Go back to your user personas to experiment with different data visualization models and experience journeys.

Step 5 - validate ideas

Select your best prototypes and take them to the field to validate with users. Allow users to play with the prototypes. Don’t defend them. Let others, not the creators, validate them. Determine which assumptions you’ve made are the most critical and identify the data that allows you to conclusively determine the correct use cases. It’s okay if your concepts are incomplete. Try and discover how users would fill the gaps and uncover hidden opportunities. If you have multiple prototypes, test which ones users are most drawn to.

Step 6 - design and test

Get started on high-fidelity designs for your data visualization experience. Apply everything you have learned to offer a data experience that will enable users to make better, faster decisions and wherever a vast supply of data is available, customize it or allow AI to continuously adapt it to their needs. If you work within an Agile team, break down a large epic into small design and development cycles that will give you a fantastic opportunity to test live prototypes with customers either in production using A/B testing or in staging. Designers can also use hi-fidelity mockups or prototyping tools to keep development costs low.

Step 7 - build and test again

Performance is critical to any data experience. Our machines have become more powerful but our consumption and production of data have equally mushroomed. Development teams and experience designers make strategic allies making technical and technology decisions that will deliver a robust data experience to users. We can easily forget that we build solutions for humans to use data to make decisions rather than to admire our cool visualizations. It’s easy to fall into the trap of creating beautiful dashboards and fancy graphs at the cost of the real value data promises to deliver. Take the time to walk in the shoes of your users, understanding the use case, craft an experience and test all assumptions to transform data visualization into a meaningful user experience.

Insight and analysis of technology and business strategy