top of page

MI Lighthouse

Project Type: Internship

Industry: Health

mi lighthouse logo

Responsibility:  Conduct qualitative research, analyzing interview data, generate wire-frame, low fidelity and high fidelity prototype 


During the pandemic, the Michigan health department faced challenges tracking vaccine data. MI Umbrella developed MI Lighthouse to aid public health, offering authorized officials visualizations for monitoring vaccine uptake, outbreak investigation, and resource allocation.
However, the platform's functionality is restricted solely to COVID-19, with limited features for vaccine campaign decision making.

The goal of this project was to:
1. Understand how users thought about their current vaccine campaign process.
2. Re-design MI Lighthouse platform integrated with other vaccine series data.

Phase 1: User Research

In total, we conducted interviews with 20 stakeholders in the public health field across Michigan. Recognizing our lack of familiarity with the daily tasks of health officers, we opted for a two-round interview approach. The first round aimed to understand their current workflow, while the second round delved deeper into identifying pain points and needs.

michigan map

User Interview R1

Tailoring our approach to different roles, we developed specific protocols for health managers and epidemiologists. In the initial round of user interviews, our focus was on understanding their current workflows to gain insights into potential pain points.

user journey

User Interview R2

We then introduced additional questions to the interview protocol to gain deeper insights into how MI Lighthouse could enhance workflow. Extracted 1000+ notes from total 20 interviews, we utilized affinity mapping to identify themes and filter potential features.

Part for the Affinity Diagram

Drag for more details


Following two rounds of interviews, we identified that the local health department officers’ daily work revolve around data management and communication, aligning with the roles of epidemiologists and health managers.


  • Local health department face challenges in accessing sufficiently granular data necessary for making informed decisions. This includes lacking geographic and demographic information for various vaccines, which hinders their ability to analyze and interpret data effectively.

  • Faculty shortage and data literacy difference among counties lead to disparity in level of data analysis.

  • Data that have time point helps local health department to track their vaccine campaign effort.

  • Local health department mainly rely on map and data comparison to check their progress and plan vaccine effort.


  • Local health department connect with various local community department such as health provider, ethnical center and school to receive data or conduct event to increase vaccination rate.

  • A part of the large vaccine effort is to connect with individuals in the community.

  • Social media is a main resources for local health department to share their data, and the decision they make related to vaccination.

  • The communication between local health department and the state faculties on data receiving is time consuming.


With that two main field, what do our user exactly need? Developing two primary user personas helps us in constructing a clearer and deeper comprehension of their needs and challenges.


Phase 2: Design

Needs prioritizing

So, we collected a bunch of wants and needs from our users. How do we sift through them to identify needs that fall within our design scope, and how do we determine which functions require a higher priority?

By creating
a MoSCoW map follow by stakeholder map, this helps us first generate all the needs from the local health department, decide what are the functions that we want to include, which also help us define the problem scope. Then based on the discussion with Project Manager and Tech lead, we use the stakeholder map to prioritize what are some function we want to include in the MPV version.


Stakeholder map

At this time, we decide to prioritize implementing key functions including color-coded maps, demographic data visualization, and PDF download capability for the MVP version as recognizing their importance.


With those precondition, we started with sketching individual elements and the overall lay out.



As we transitioned to wire-framing, we prioritized the map and dashboard pages, focusing on a select few options favored by the project team. This approach allowed us to brainstorm various layouts that could be developed into low-fidelity prototypes.


Things we want:The ability to show options to the user related to their preferencesInforming the user about how to navigate between regions

Things we noticed & liked:

  • bookmark tabs to shift between the different vaccines

  • users saving preferences in some way (views they enjoy)

  • customizing view to type of user (local epidemiologist vs state epi vs school administrator)

  • Search function for anything (person, county, tract, boundary)

  • “whats going on this county” - recent updates in their areas regarding significant changes in vaccine/outbreaks

Feature Decision for landing page:

  • download data (individual-level data)

  • provide recent, granular updates

  • default to local county toggle between tract/school and between Flu/Covid/MMR

  • search functionality on landing page

  • Case and vaccine data showed side-by-side

Low Fidelity

As moving forward, we create more detailed content on each page and even a little bit high fidelity. At this time, we main focus on two lay out, one is with the top bar with filters on the right side bar and the other have a fixed sidebar on the left.

pros:top bar is similar as version 1, less learning curve for users

cons:two pages are not consistent which against the heuristic rule

low fi

pros:pages are more consistent with the fixed side bar

cons:the information hierarchy is not clear enough on the side bar

Upon developing both options, we found it challenging to decide which style to proceed with for following reasons:

  • Maintaining consistency in data analysis across tab sizes poses a challenge due to varying levels of analysis required for the map and data analysis pages.

  • However, certain elements must remain fixed , such as the switch between the map and data pages, or toggling between different vaccines.

So, instead of committing to either design, we opted to take some time to explore other works related to maps and dashboards to gain additional insights and see if we can make adjustment to come up with a new style and move forward to high-fidelity.

High Fidelity(Still in progress)

high fi map

The current design of the map page integrates elements from previous versions, featuring a fixed left bar for switching among metrics , map and other navigations we might include in later version.
Meanwhile, filters have been relocated to the top bar, as they are not displayed on the metrics page.
This layout ensures that users primarily interact with the right panel, utilizing the left menu bar solely for page navigation or accessing account settings.

high fi metrics

The metrics pages contain a difference in different size of screen. The reason we include an overview in the monitor screen, which is the size most local health department officer interact with is that it can provide a quick idea of how the trends look like in that graph. But we still need more testing to verify whether this difference will lead to confusion.

What's next? & What have I learned?

We are still working on iterating the high fidelity pages and waited to conduct user testing for feedback!  During this design process, here are some points I noticed:

Balance users need and real word situation

As a designer, I recognize that the real world is constantly evolving, and users may express a multitude of needs during interviews. Even though these insights are valuable, it's essential to filter and prioritize them based on the project scope and timeline to focus on those that will have the greatest impact. Trying to design of everyone might lead to design for no one. No single product can address every user need comprehensively, delving deeper into a specific scope allows for a more profound understanding and enables the creation of solutions that resonate more deeply with users.

Think out of the scope

At the beginning of the design stage, our team engages in a "crazy 8" exercise, involving all members regardless of their roles. This exercise offers insights into diverse design perspectives, especially those from developers, who present layouts distinct from designers'. I've personally found it challenging to break out from a specific scope, resulting in similar designs. However, hearing input from individuals not directly involved in the project, or from those with different roles, might be beneficial. Their perspectives enable us to break out of the scope and approach design challenges from fresh angles.

bottom of page