Princeton research community, including faculty, PhD students, research assistants, lab managers, data analysts etc.
Research data management - giving data proper labeling, storage, and access at every stage of research
A comprehensive data management ecosystem:
- centralized platform for diverse projects and data storage
- streamlined processes for metadata labeling
- supporting long-term sustainable research and data growth
At Princeton, researchers store their data across multiple platforms and clusters, and oftentimes lacking the awareness or motivation to label it properly, making data hard to find and reuse.
What’s worse, with busy schedules and varying technical skills, researchers face challenges managing data efficiently.
To explore the problem space and understand users' pain points, we initiated the research and engaged directly with over 100 researchers, leading to the ideation of 200+ ideas, which we then refined into tangible solutions.
Data management practices are scattered and unstandardized, our survey revealed 46+ distinct tools used for managing research data.
Researchers organize their data in very different ways, which leads to confusion when sharing data, causing ineffective inter- and intra- team collaboration.
Researchers face technical issues related to data storage, scalability, and poor usability of existing tools and platforms.
In traditional workflows, researchers switch between multiple data management tools either due to requirements, limited storage space, or because certain tools are commonly used by their disciplines.
TigerData system offers a centralized, customizable space for Princeton research community to manage all their projects and data, including those with massive data.
TigerData serves as a centralized hub for all projects and data, allowing users to track storage usage, manage data using quick actions, and more.
• Save time from switching between tools
• Quick action reduces navigation time
TigerData Discovery enables searches for active research data in Princeton’s repository using metadata and relevancy level, promoting data sharing and reuse.
• Reduce the effort collecting all data from scratch
• Discover potential collaborators
To simplify and encourage metadata tagging, this intuitive file tagging board enables researchers to apply schemas and metadata to thousands of files with just a few clicks.
• Less redundant manual tags input
• Save time by applying pre-defined schemas
The embedded query builder uses intuitive visual components that allows those with less coding experience to create custom filters for their data with minimal learning.
• User-friendly for non-technical users
• Build, save, and apply directly next time
Based on the issues uncovered in our initial research, we used a unique ideation method, where we drew a random combination from our user group, problem and technology, and design for it. We think beyond current limitations and sketched over 200 seemingly crazy ideas.
How can we bring those ideas back to reality? We first validated user needs through speed-dating, then used methods like “user enactments” to envision potential solutions. Finally, we refined the concept into something feasible for implementation on TigerData.
In a collaboration workshop with the TigerData team, we talked through all potential solutions and mapped them on a priority matrix based on their alignment with business objectives and technical feasibility for the team to implement.
To ensure optimal user experience, our team conducted two rounds of usability testing. By coordinating with TigerData team, we were able to engage actual potential users from various departments, which led to some major improvements of the design.
Issue 01: Users frequently interact with projects, but need extra scrolling to access them.
Solution: Prioritize project overview to make it central focus and position other activities to the side to better align with user behaviors.
Issue 02: Users were unclear about their roles within different projects and overlooked the role categories in the tabs.
Solution: Consolidate all approved projects into a single tab and clearly display user roles for each project to provide transparent information.
Issue 03: Users felt uncertain about what to expect at each step when creating a project.
Solution: Add descriptions and status indicators to the progress bar to provide clear guidance on “to-dos” in each step.
Issue 04: Users were unsure of how much storage they needed, oftentimes went directly to select the maximum available amount.
Solution: Provide multiple storage options with clear description to help users choose the optimal amount based on their specific needs.
Issue 05: Users felt hesitant about the terminology and how the query logic work.
Solution: Restructure the visual hierarchy of Boolean logic and adopt familiar terms used in other tools to make the process more intuitive.
Issue 06: Some users preferred coding queries directly rather than using step by step visual representation.
Solution: Introduce an option for coding interface that allows technical users to build queries in their preferred format.
Issue 07: Users were confused when adding one metadata tag led to multiple tags being added.
Solution: Use arrows to visually represent the parent-child relationships between tags to make the hierarchy clearer.
Issue 08: Some users mistook the expand button (the “+” sign) for an “add” button.
Solution: Replace the “+” and “-” signs with commonly recognized arrow icons to reduce confusion and better represent expand/collapse.
TigerData is currently in its MVP stage, but we envision it as a living, evolving tool that adapts to the changing needs of its users. Therefore, we value the sustainability of design while designing the product.
As our team worked on different part of the designs, we approached the design of components very differently. We recognized the need for a well-documented design system to align our work and ensure TigerData team can reuse and build upon the design after we handoff all the deliverables.
In two weeks, we built a design system for TigerData using Atomic Design Methodology, consisting of 5 tokens and 15 component sets, leveraging the use of properties to maximize efficiency and flexibility.
I created a strategic roadmap, envisioning how TigerData may advance from Day 1 to the next 5 years, validated by the TigerData team. Each phase includes specific goals, features, and metrics to measure success and track adoption.
This problem place brought many opportunities, but designing an institution-wise tool also came with many constraints. I appreciated how my team explored “risky ideas” early on, which sparked valuable insights and guided us as we transforming bold concepts into tangible designs.
Working with a large stakeholders group is difficult, especially many of them are leaders in their respective fields. I found that the more we demonstrated and showed genuine appreciation for their insights, the more open they became in sharing their feelings and thoughts. Successful product development needs to build on mutual trust and respect.
Working on an early-stage product like this taught me that having a grand vision is important, but it’s equally crucial to pay attention to the smallest details like a button, or how data is displayed. These micro-level adjustments helped us deliver a more usable and accessible product.
