NASA's Eclipse Soundscapes' Data Analyst page relies on volunteers to analyze eclipse audio data, with a mission to make that experience accessible to visually impaired participants. However, the training page buries that role behind six confusing, text-heavy sections — undermining the accessibility the project is built around.
I restructured the Data Analyst page's information architecture to create a clear user journey and prioritize accessibility. My team and I conducted preliminary research through competitive analysis and site mapping. We then prioritized changes through user research with activities such as card sorting and usability testing.
Eclipse Soundscapes is a NASA funded citizen science project that invites the public to explore how solar eclipses impact life on Earth. Their mission is to allow visually impaired individuals to experience the eclipse through sound and touch. This project was a semester-long client engagement with NASA's Eclipse Soundscapes team through ARISA Lab, scoped specifically to the Data Analyst training page.
How might we improve upon the information architecture of the current Data Analyst Role page in order to ensure clarity and accessibility for users to better accomplish the task of analyzing audio data?
We had 6 users, who are young adults that have no experience with citizen science projects, navigate through the original website. In these usability testing interviews, we tracked both quantitative and qualitative data in order to find out: How is the site currently working?
"I would like you to explore this web page with the goal of analyzing some audio data and input your analysis into a form. Go through the web page and try to achieve this."
This task took an average of 6 clicks with a 1:43 completion time.
The following are screens of the original site.
Afterwards, we conducted qualitative interviews and synthesized the following:
We analyzed three other role-training platforms built for the shared goal of helping new users understand and retain unfamiliar material, asking the question: How do other companies do it?
Eclipse Soundscapes |
AWS Certified Cloud Practitioner |
Microsoft Azure Training |
Google Cloud Training |
|
|---|---|---|---|---|
| Successes | Steps are laid out clearly | Clear growth pathways, resources centralized | Tabs cut down on scrolling | Visually clean, clear tabs |
| Pain Points | Too much text; certification hard to find | Training buried several pages deep | Too many pages; learning paths aren't visual | Certification tab easy to miss |
| Takeaway | Separate training content from role info; cut scrolling | Use one page as a launch point, not a maze | Use tabs/bookmarks to reduce scrolling | Balance visuals and text without adding pages |
We conducted a card sorting activity with 3 users in order to gain a better understanding of users' mental models.
From this, we learned that reorganizing the site meant prioritizing users' goals of understanding, training, and completing data analysis — and cutting down on the number of tabs.
Map of original information architecture.
Prospected map after informed research.
Section Changes
Based on our research, we translated user frustrations into redesign goals that emphasize the user journey.
- To have a better balance between text and visuals
- To navigate more intuitively and easily
- To spend little to no time finding links and information for task completion
- Site is built on WordPress — no custom framework, limited to what the CMS and available plugins support.
- Site must remain in light mode only — no dark mode toggle.
- Survey platform is Google Forms, embedded directly into pages.
- Scope limited to the Data Analyst role page — not a full-site redesign.
In our redesign, we were required to use the following branding colors and accessibility guidelines from NASA Eclipse Soundscapes.
Our team each created a low fidelity wireframe of our potential redesign. We referenced these when collaborating on our final prototype.
Every core design decision traces back to a specific research finding. Collapsing 6 sections into 4 targeted the 6-click / 1:43 average it took users to find the analysis form. The visual design addressed the biggest complaint directly: too much text, not enough visuals. Infinite scroll answered two problems at once — sighted users losing their place while scrolling, and NASA's own brief for VoiceOver compatibility.
- Task completion time and click count on the analyze audio data flow, benchmarked against the 6-click / 1:43 baseline
- Drop-off rate by section — to confirm four sections actually reduce abandonment rather than just relocating it
- Accessibility (WCAG) conformance score via WAVE tool, plus a manual VoiceOver walkthrough
- Volunteer return rate — whether people who complete one audio analysis come back for a second
Testing what we could actually test. NASA's brief named blind and low-vision volunteers directly, but recruiting people who use screen readers daily wasn't realistic on a five-week class timeline. Every accessibility decision — VoiceOver compatibility, contrast, font sizing — came from WCAG guidelines and best practice, not from watching someone actually navigate the page with a screen reader.
If I had more time… I'd close that gap directly — recruiting blind or low-vision volunteers and re-testing specifically for screen-reader compatibility, instead of assuming best practices translate cleanly to lived experience.













