Uncovering what researchers need from a genomic data platform
MY ROLE
Product designer
TIMELINE
April 2023 - December 2023
TEAM
CHALLENGE
Feature growth outpaced user experience
NCBI Virus is a free NIH platform used by thousands of global scientists to search, analyze, and download viral genomic data for public health and vaccine development.
Following its 2018 launch, rapid feature expansion—accelerated by urgent COVID-19 data demands—left the platform organized around technical features rather than actual user workflows.
PROBLEM DISCOVERY
Analytics revealed a bypassed homepage, misordered filters, and a hidden dashboard
I audited the entire platform using usability heuristics and in-house analytics from March 2023.
01
Homepage: 15x fewer visits than the Results Table.
02
Navigation and tools: ordered by history, not by use.
03
Filters and columns: 24 filters in no logical order, with high-use ones buried mid-list.
04
Dashboard: reachable only from the Results Table, with 30x fewer visits.
While data highlighted where the structure was failing, it couldn't explain why or for whom.
Home page with Dashboard 1: the top action is buried
15x fewer visits than Results Table
Most-clicked "Search by virus" (#1) has no more weight than BLAST search and sits below COVID alerts
Results Table: filters aren't ordered by use
Most visited page
The second most-used filter, Geographic region (#2), sits 11th in the list
Dashboard 2: a separate page that was hard to find
30x fewer visits than Results Table
Reachable only from the Results Table; missing from homepage and global navigation
High-demand features are buried
Heuristic evaluation + usage analytics. Numbers 1–3 show the most-used features on each page (1 = highest)
RESEARCH GOALS
Targeting who our users are, what they do, how they work, and what they lack
To understand the reasons behind the numbers, I set up a user research program with four goals:
01
Who
Who uses NCBI Virus?
Has that changed since we built it?
02
What
What do researchers try to do?
Do they succeed?
03
How
How do their workflows run?
Where do they break?
04
Gaps
Which features are missing?
Which could be retired?
METHODS AND PARTICIPANTS
Using mixed-method user research to fill the analytics gap
Usage data only tells you what is happening. To understand why and how, I combined four distinct research methods—balancing quantitative scale with qualitative depth.
Research methods decision tree
Participants
All participants were unpaid NCBI Virus users. The survey and comment cards were open to anyone, since we wanted to learn who our users are. Interviews were with scientists who had cited NCBI Virus in their papers or used it regularly.
User interview notes
RESEARCH ANALYSIS AND FINDINGS
Scientists did most of their work in the Results Table, struggled with its navigation and filters, and rarely found the dashboard
How I analyzed it:
I combined statistics and affinity mapping, counted how often each theme came up across all sources, mapped users' workflows, and prioritized the findings into recommendations.
Quantitative
Statistics on the survey, comment cards and emails
Qualitative
Affinity mapping of interviews, usability tests and open answers
Synthesis
Counted each theme across all sources and grouped the findings into four categories
Used elements
Liked
Struggled with
Asked for
What I found:
01
Our largest user groups are bioinformatics researchers and master's degree students
42% researchers, 41% students. 80% returning, 78% arrived from a search engine. Most were satisfied (72%); the problems were specific.
02
Two main workflows run through the Results Table
Scientists used the platform in two ways: quick data discovery (see what data exists, filter, download) and dataset analysis (search for a virus, filter, visualize, download for phylogenetic analysis). Both ran through the Results Table, which explains its heavy use: it drew almost twice as many comments as the visual tools, more than any other part of the platform, so its problems affected every user.
03
Filters that were hard to find and hard to read
Scientists liked the core filters (Virus, Collection date, Completeness, Host), but found them hard to locate and couldn't tell what some filters did from their names alone.
04
People wanted visual tools, but couldn't find them
The visual tools were the second most discussed part of the platform. Scientists liked the map, time sliders, and taxonomy and host widgets, yet many didn't know the dashboard existed, which explains why so few reached it.
Slides from the research readout presented to the team
RECOMMENDATIONS
A three-stage roadmap: quick homepage wins, followed by the Results Table and dashboards
I ordered the findings by effort and value, so the team could ship early while planning the harder work.
01
Home page
Low effort (no backend work)
Every visitor sees it first
02
Highest value
Results Table
Medium effort (research first)
Workflows run through it
03
Dashboards
High effort (technical)
Hardest to build, so last
01
Simplified the homepage without backend changes
Made search the main call to action
Cut navigation to the core workflows
Added help links where scientists get stuck
02
Researched filter use before redesigning the Results Table
Studied how scientists name, group, and use filters
Reorganized filters and columns around the findings
03
Unified the dashboards and connected them to the Results Table
Merged two dashboards into one
Synced filters with the Results Table
LEARNINGS
Sharing actionable findings, leading with priorities, and planning research in phases
Through this project, key insights emerged regarding the communication of findings alongside its execution.
01
Lead with clear, actionable guidance
Rather than sharing every detail, give the team concise recommendations with brief reasoning and a clear order of work. Keep the full evidence available for anyone who wants to look deeper.
02
Document every phase
Research often happens in stages, with pauses for more urgent work. Thorough documentation at each stage lets the work pick up where it left off and makes follow-up studies possible.
OTHER WORK
Guiding researchers to critical genomic data through a homepage redesign
Accelerating data discovery providing a clear call to action