Robin Stewart: portfolio
icon

Sculpting Data

Vizable for iPad & iPhone

Vizable is a mobile app whose design arose from the belief that statistics and data analysis could be dramatically easier to use. I invented and prototyped the core user interfaces and led the app's design as the team quickly grew to 15 members.

Taxonomy of data analysis use cases (blurred)

Research

I spent about six months learning everything I could about the questions people most want to answer with their data. I read books, research papers, and even a Ph.D. thesis, and repeatedly presented my findings and sketches to colleagues with deep customer experience, who helped me refine these ideas. As my understanding grew, I iteratively developed a taxonomy of data analysis use cases, aiming to uncover the simplest information architecture I possibly could. One version is shown at right.

Prototyping

Once I was confident in the high-level taxonomy, I began developing prototypes to explore and demonstrate my vision for highly interactive, direct-manipulation data analysis. I usually started by sketching ideas in OmniGraffle and then moved to Xcode/Objective-C to build functioning, prototype iPad apps.

For example, one of my earliest frame-by-frame sketches explored the idea of removing a category of data by "pulling" a bar out of a bar chart using your finger:

The stretchy, elastic concept helped the app feel tactile and playful. However, when I converted it into a live prototype for testing, it became clear that changing the length of the bar as the user drags their finger looks too much like an action that alters the data (rather than just removing that bar's data from the view). So I switched to a transition that keeps the bar at its original size as it gets pulled away (along with some other fine-tuning):

To better understand and demonstrate the full power of this approach, I needed a prototype that was advanced enough that people could use it to ask real questions about real data. To achieve this, I combined several previous prototypes, programmed a rudimentary database back-end, and devised a technical architecture that supports smooth interpolation between any two database queries as the user drags their finger from one state to another. I call this interaction approach "sculpting data" because it takes inspiration from the intuitive, hands-on process of shaping clay.

Among other things, this functionality allowed us to discover an emergent property of the design: a whole new way to answer comparative “what-if” questions by simply sliding a finger forward and back. This technique is demonstrated toward the end of the following video, where I analyze fictional sales data from a chain of coffee shops.

The analytic power of even this relatively simple prototype was enough to convince the executive team that the project was worth pursuing in earnest.

Development

Over the following year and a half, I helped hire new team members and worked to explain the research, design and engineering behind my prototypes. I collaborated with engineers as they filled in missing pieces of the prototypes and built a more powerful database and many other features. I mentored designers working on new features and I advised design agencies who developed the app's visual design and branding. Throughout, I met frequently with executives and project managers to brainstorm and prioritize product development.

Meanwhile, I continued to create new prototypes, exploring novel interaction techniques for time series analysis, geospatial analysis, data distributions, statistical tests, and more. For example, the prototype below lets you continue to zoom into a timeline beyond the summarized data and directly into individual underlying records:

Another concept lets you use an unpinch gesture to "zoom into" a bar chart. It incrementally breaks down the original categories by adding a second field, providing a more detailed analysis:

Some of the most interesting prototypes are still under NDA, so I can't share them publicly. I'm the inventor of nine patents that Tableau has filed on this work.

Results

We released the first version of Vizable in October 2015. It was profiled by hundreds of news outlets and featured in the Apple app store for many weeks.

“Vizable… converts spreadsheets into easily manipulatable charts and graphs, making it dead simple for small businesses or individual users to parse through and understand large amounts of data.” -Meg Miller, Fast Company Design
“This is the first tool that helps me feel the presence of the data I create… This is incredibly impressive work.” -Jason Preston, founder, DENT conference

Since then, users of all ages have been analyzing data in every imaginable domain, from budgeting to popular culture to science to sports to politics to business.

“In the few months I have been using Vizable, I have learned more about my family's business than I probably have in the last five years. … You can change, rotate, and see the data in different ways with minimal effort. So if you have a question, you can answer it immediately.”
-Alejandro Palaez, consultant
“Easy to use and absolutely amazing. A great way to visualize data and present information quickly and easily.” -Mariogeronimo, app store review
“I needed a way to analyze trends… What I stumbled on was this jewel of an app. It did WAY beyond what I had even dared to hope for… A truly magnificent app! Thanks for making data analysis so fun and useful!”
-Oopiwapoc, app store review

Further Reading

Robin Stewart. Tactile Data Analysis. Blog post for the Vizable community.

Robin Stewart. Data Worlds: The Model Behind our New App, Vizable. Tableau Conference 2015.

Vizable product website: vizable.tableau.com