Program Analytics | HHMI
I designed, developed, and implemented Program Analytics to longitudinally track and analyze the professional trajectories of all HHMI program participants and alumni.
“Program Analytics was my response to a fundamental problem: how do we turn scattered spreadsheets and disconnected systems into a unified infrastructure that tells the full story of our impact?”
Program Analytics: Building a System to Track People, Not Just Numbers
HHMI programs have supported thousands of scientists, educators, and students over the years—but we had no reliable way to keep track of them. Most data lived in spreadsheets, scattered across departments, disconnected from our evolving goals and commitments. There was no shared understanding of who counted as a participant, no consistent way to update information, and no central system to connect the dots. The result? Lost institutional memory, missed opportunities for collaboration, and a limited ability to tell the full story of our impact.
I led an initiative to change that.
Program Analytics was my response to this challenge—a project designed to create a scalable, integrated system for tracking program participants and alumni over time. But it wasn’t just about building a database. It was about designing a culture of data that could evolve with our programs, reflect the people we serve, and support strategic decision-making across the Institute.
As part of this project, I authored a detailed business case to build internal alignment and secure support for Program Analytics. The document outlines the rationale, challenges, proposed framework, and implementation strategy—including a two-phase approach to metadata development and CRM adoption.
📄 Read the business case
The work began with people. I met with colleagues across departments—from science, operations, human resources, educators to systems analysts—to ask a deceptively simple question: Who do we serve, and how do we describe them? From those conversations, I developed a shared metadata framework—a common language for defining populations across their career stages, grounded in national standards and institutional knowledge. This metadata became the backbone of the system.
🏷️ Browse the metadata document
I then turned to the technology. I mapped out a user-centered interface that would allow program teams to enter, update, and analyze participant data with ease and consistency. But translating a high-level vision into something actionable required more than just technical specs—it required a clear, visual language. That’s where wireframing became essential. I created detailed, clickable mockups of the system’s interface, including the sign-in experience, dashboard views, data entry forms, and population filters. These wireframes served as a bridge between strategy and execution, helping business stakeholders understand how their needs would be addressed in the design, and enabling the visualization team to see how interface elements could support key user actions and data workflows.
In envisioning the dashboard, I focused on tying insights to the questions that mattered most to each program—whether about alumni career outcomes, diversity goals, or program effectiveness. The wireframes helped surface those needs early, ensuring the system would not just display data, but help users make sense of it. This visual-first approach also set the stage for downstream development in Tableau, Qualtrics, and Microsoft Dynamics, giving each technical team a shared blueprint to work from.
🖼️ The wire framing is here
As the scope of Program Analytics expanded, it became clear that we needed a durable data structure—one that could represent the complexity of scientific careers, programs, and institutional affiliations without flattening them into spreadsheets or single records. To make that architecture visible and actionable, I developed an Entity-Relationship (ER) diagram that served as a critical blueprint for both technical development and stakeholder alignment.
This diagram maps the key relationships among participants, programs, grants, institutions, and career outcomes. Each object—such as Contact (Scientist), Career/Occupation/Job, and Organization—is connected via lookups and join tables that support longitudinal analysis. For example, one scientist may participate in multiple programs, hold several positions over time, and be associated with different industries or classification systems (e.g., NAICS, AAUDE, SOC, CNGLS).
When fully implemented, Program Analytics would make it possible to track more than 6,000 alumni, connect participant journeys across decades, and align our data practices with the mission and aspirations of the Institute. In short, this project showed how data infrastructure can be about more than efficiency—it can be about memory and the relationships we build over time.
