Statistical software is generally engineered for statisticians. As niche software, it is technical and complicated intended for a user who is already very familiar with the tools of statistical analysis. Highly specialized software is often, if not always, poorly designed from a user interface perspective. The client, MoreSteam, has an ambitious goal to make statistical analysis more user friendly, experimental and sharable. Additionally, MoreSteam chose to exploit the latest web standards to create a statistical software package that exists purely as a web application.
EngineRoom is a new web-based stats package based on the firm's legacy product of the same name, which exists as a powerful, though cumbersome, plugin for Microsoft Excel. Internally, the MoreSteam team has the expertise to develop and deploy such a product, but realized the importance of good UI/UX in order to reach a broader audience of those who are new to stats, like students, or those who occasionally require statistical analysis but do not have a need to master it. To help in this area, MoreSteam added Andre Murnieks of ACM Interactive as a design consultant embedded with the development team for an intensive series of sessions to brainstorm and produce a model for the new web interface.
As the team progressed through the design phase, everyone became very receptive to principles of not only good UI/UX principles, but also the nuances of overall visual communication design as evidenced in the use of typography, iconography, a restrictive color palette, motions and hierarchical layering of interface elements. The primary interaction of the new EngineRoom is centered around the flexible hexagon acting as a visual abstraction of statistical methods and data types. The user may freely experiment with analyses by dragging and dropping different data types onto various statistical methods creating honeycomb clusters representing a study. When a cluster has the requisite data, the appropriate output or visualization is dynamically created.
Since each cluster is instantly saved, the real power of EngineRoom for the novice is that multiple analyses can be quickly generated, reviewed, saved or later discarded. The speed at which the user can explore the various statistical methods not only encourages experimentation but also builds confidence with guidance and explanation along the way. Statistics is a field requiring education and skill, but for those who have some knowledge, but not expertise, EngineRoom's intuitive interface eases the burden and makes big data more approachable and manageable.
research & subject matter experts
The original design team that participated in the "deep dive" of imagining a radically graphic of approaching statistics was made of MoreSteams senior leadership and expertise with advanced degrees in finance, operations research, statistics (naturally), computer science, industrial systems engineering, nuclear engineering and physics. While most, if not all, of the team have moderate to extensive knowledge of statistics ("Lean Six Sigma Black Belt"), the visual communication designer did not. To facilitate a dialog between the experts and designer, a series of personas with realistic use cases were developed to guide UI/UX decisions. The nature of the personas tasks and goals were a valuable means through which the designer could understand the technical requirements, while also providing a narrative through which to communicate UI/UX principles and rationale.
During the initial design phase, front-end and back-end developers were able to create proofs of concepts for various UI/UX ideas—almost as fast as the ideas were sketched.
ethnography, testing & distribution
Once the EngineRoom site was live in both beta and initial releases, MoreSteam continued to refine the application directly observing the user's trials and tribulations with the UX/UI. Because EngineRoom is entirely web-based software, reacting to user input and issues is a quick and continuous process.
EngineRoom is a GUI front-end for R, a industry standard language and environment for statistical computing and graphics. It is a GNU project developed at Bell Laboratories (formerly AT&T, now Lucent Technologies) by John Chambers and colleagues. R is an extensible, open-source data manipulation and calculation platform.