Case Study: Searchlight/Polaris Data Search

The Challenge

The National Snow and Ice Data Center (NSIDC) archives and distributes a vast amount of snow and ice related satellite data from NASA and other federal agencies. While most of this data was accessible through their website, there were a plethora of different interfaces and portals to choose from. Users had to be “in the know” about which specific grant or project was responsible for specific data sets in order to find them. This often left people confused and irritated. It also left the data buried where it’s value couldn’t be realized.

The ambitious goal of the Searchlight (later renamed Polaris) project was to create a single user interface where all NSIDC data would be not only searchable by geographic constraints and data types, but also to provide on-the-fly imagery and data analysis. The stakeholders and Product Manager envisioned a “one stop shop” for their data that was clean, intuitive, and easy-to-use.

The Approach

I served as the UI/UX Lead for this project, and worked closely with the Product Manager and a talented team of of Software Engineers, Category Experts, Scientists, and Writers to bring the new vision of data search to life. From the beginning, I successfully advocated for the importance of building UX design work into the Agile/Scrum process.

UX design sprints ran ahead, or sometimes in parallel, to the very complex technical side of the project. First, I worked with the CMEs, Scientists, and Writers to understand the audience and create a suite of Personas. These were very helpful for keeping the team and stakeholders focused on the value being delivered to our users.

At every stage of the project, there was lots of Sketching and working with Paper Prototypes in order to test concepts and get feedback. I put new iterations of interface components in front of NSIDC Scientists on a daily basis, and brought their invaluable perspective back to the team. This kept the project moving in the right direction, making certain we were building the right things.

Sketch of Faceted Search for Searchlight/Polaris
Sketch of Faceted Lists for Searchlight/Polaris

Sketch of Searchlight/Polaris User Interface Elements
Sketch of Searchlight/Polaris User Interface Elements

Following the Agile principle of working software over documentation, there was a strong push to build something fast. Therefore, there was very little lag time between Sketches and Wireframes and code development. For this reason, the validation we’d demonstrated with the low fidelity deliverables and testing was incredibly important.

Wireframe of Searchlight/Polaris SERP and Data Queue
Wireframe of Searchlight/Polaris SERP and Data Queue

Later in the project, when the product was approaching being “finished,” at my suggestion we initiated a series of focused Usability Studies. This had never been done before at NSIDC, so I used webinars and other resources to train our diverse teams on the purpose and techniques of Usability Testing. The team and stakeholders were very excited about the process.

We invited a variety of Scientists to participate in the Usability studies, and for the second round we also had some NASA stakeholders on-site to observe the testing. The first study revealed some subtleties of language and mental models that we’d missed, and we were able to correct these before the second study.

The second study, from a certain perspective, was a disaster. The application servers couldn’t handle more than a few concurrent users, and the application slowed to crawl and finally crashed. The NASA folks basically caught us with our pants down. However, their overall impression was surprisingly one of professionalism and rigor based on the usability testing process. This was an incredibly important lesson for the team about the importance of planning for application speed and performance, as well as the very real value of usability testing.

The Results

From the technical side, the Searchlight/Polaris project was a major achievement, and from the User Experience perspective the final product was a dramatic improvement. There was now ONE prominently featured data search interface for ALL of NSIDC’s data, and it was simple enough for anyone to use. It was therefore easy to find, and could filter data based on things like geographic parameters, date and time constraints, data instrument type, and many other variables. The system also allowed for browsing data imagery, creating different types of data visualizations, and re-projection of maps to different types, all of which were huge wins.

Mockup of the Searchlight/PolarisHome screen
One of the “final” mockups of the Searchlight/Polaris Home screen

Scientists from NSIDC and other research organizations, and the funding stakeholders, were very impressed and grateful to the team for their work. Our friends from NASA were actually delighted with the end result, and repeatedly applauded the process that delivered such a product. The team came away from the project feeling like we’d done something good, made something better, and made our stakeholders happy.

Learn more about my work for NSIDC.

This entry was posted in Case Studies, Front-End Development, Interaction Design, User Experience, User Interface, Visual Design, Web Applications and tagged , , , , , , . Bookmark the permalink.