I worked with Kinsa Health to build a publicly accessible news source for contagious illness, built around a unique dataset from Kinsa's network of app-connected smart thermometers.
HealthWeather helps Kinsa demonstrate the value of their real-time disease monitoring system in order to partner with local government clients to better prepare for future outbreaks.
HealthWeather offers a real-time view of disease risk in your area - much the same way a weather website offers a forecast to make travel planning and group activities easier.
Healthweather was an opportunity to build a unified design system for presenting data concepts and a chance to explore how Kinsa expressed its brand voice across products.
Kinsa is a healthcare product company on a mission to change the way we monitor contagious illness by bringing the same sort of early warning and public awareness to the spread of disease that we have for so many other persistent threats in our communities.
I contributed to this project as a Data Product Designer. I was reponsible for executing all visual designs in this case study.
The world of data visualization around COVID-19 evolved daily, so understanding how HealthWeather fit into this landscape was essential.
Our team and I reviewed many news sources and data visualization outlets to better understand how the risk of COVID-19 was being made accessible to the public online.
In its early stages, HealthWeather relied on written analysis to provide insights on the spread of illness.
I worked collaboratively with other designers to wireframe and explore the integration of written analysis into HealthWeather's homepage while the data science team honed models for generating automated insights.
As a exploration tool, early iterations of HealthWeather focused on mapping unusual fever levels.
Iterations at this stage allowed users to directly observe fever data from Kinsa's network of smart thermometers. By mapping fever levels that were unusually high for a particular time of year, users could see areas of increased likelihood for COVID-19 outbreaks.
We also explored HealthWeather as a means to support the general public in better understanding the mechanics of the pandemic from a data-driven perspective.
Interface concepts like this focused on showing users the connection between elevated illness levels and interventions by state and local governments. Using data-driven design tools, I helped explore branded visualization concepts in high-fidelity for team discussion and brainstorming.
Wherever possible, I opted to create high-fidelity visual design around accurate data to elevate brainstorming and team conversation.
I used a combination of raw data, Tableau visualizations, SVG charts, and Figma plugins to make sure that teams were seeing the most realistic visualizations possible - and that the engineering team had precisely styled chart designs (including typographic/color styles that matched their CSS variable system) to make development smooth and efficient.
In our initial work, we focused on supporting a more technical user group, such as epidemiologists and journalists who would understand terms like "atypical illness" and "%ILI." As our work progressed and the team recieved more input from users, we began to look for ways to simplify HealthWeather's communication in order to be more accessible to a general audience.
As the HealthWeather concept grew, there was an increasing need to represent multiple variables across time and geographic location.
Making use of as many component-based features and plugins as I could find, I used design tools like Figma to make exploratory "design sandboxes" to help show how small design changes to color, iconography, and layout could have far-reaching effects on how users perceive warnings and explanations in different displays.
While many of my initial explorations into icon systems and color schemes were exciting from a design perspective, we increasingly received feedback that our interfaces needed to quickly answer the question "how worried should I be?"
Paper sketching helps me quickly explore communication ideas that are highly customized to the specific problem I'm trying to solve, allowing unexpected connections to emerge.
Although I strongly support letting digital tools speed up the process of layout design and to increase consistency across interfaces, there are times (espeically in data visualization) where the messiness of hand-drawn sketches can be a big advantage in uncovering visual metaphors and connections that might otherwise stay hidden.
As Kinsa's data model evolved, HealthWeather progressed past a visualization of fever data into the possibility of a unique early warning system for COVID-19.
HealthWeather evolved at a rapid pace from one information display to another, and I kept pace with its evolution by regularly communicating with stakeholders on many different facets of the initiative - data scientists, product management, leadership, and more.
Because HealthWeather took on the conceptual footprint of a news site, I made sure not to assume that all users would be accessing the visualizations from a large screen.
Even though the days have hopefully passed when responsive design was considered an advanced feature, it's still worth mentioning that I thought through all visualizations on the HealthWeather site from a flexible perspective that went beyond set breakpoints. Within reason, HealthWeather works on any screen.
However, when I watched users parsing detailed citations for our data sources, the complexity of the information seemed to increase skepticism. Counterintuitively, when we showed designs that focused on risk using simplified visualizations, users focused more on the value of the information and asked fewer questions about its source.
As the national conversation around COVID-19 shifted, I helped transition and set up design components for future features and development.
Even though much of the design process thus far had focused on specific, COVID-19 visualizations, I made a careful effort throughout the entire process to make layouts, structures, and visual styles flexible. Using variables, design components, and responsive layouts, I streamlined the process of reimagining the interface.
Kinsa took action on an unprecedented opportunity to show the public how an illness monitoring system could become a part of daily routine, just like checking a weather report online.
Coordinated efforts from many different teams at Kinsa resulted in over two million users in HealthWeather's first nine months as a source for illness information. Additional metrics revealed how the teams' strategic thinking inspired users to spend more time on the site and bounce less often. Further, The Centers for Disease Control and Prevention (CDC) integrated HealthWeather visualizations into their internal reporting.