Prototyping with data: Visualizing environmental triggers of asthma
By Emily Saltz, Michelina Campanella, and Jeffrey Chou
CoPilot is a service we’re designing to provide children, their parents, and school caregivers with actionable information about asthma triggers in the environment to answer one simple question: is it safe for an asthmatic child to play outside on a given day? Through primary research, we learned that both children and their caretakers rely on careful planning in order to coordinate weekly activities: if a dangerous level of pollen is expected on a certain day and a child needs to stay inside, the parent and teacher would prefer to plan for this sooner rather than later. As a result, we’re hoping to design our service to notify parents, teachers, and children about predicted environmental conditions for each upcoming week.
Searching for Data on Asthma Triggers
To prove out the feasibility of generating localized weekly notifications based on changing environmental conditions related to asthma, we searched for data sources that could be used as the foundation of the weekly forecast.
We began with a list of nine environmental triggers of asthma from the CDC, and tried to narrow down which triggers could be used as the foundation for notifications within the service:
3. Secondhand Smoke
4. Dust Mites
7. Nitrogen Dioxide
8. Outdoor air pollution
9. Chemical Irritants
Because several triggers, such as secondhand smoke, dust mites, pets, cockroaches and chemical irritants appear appear within narrow location ranges, such as inside apartment buildings, we weren’t able to find any robust datasets that conformed to national standards. As a result, we focused on the areas of air quality, pollen, and mold.
Air Quality Index (AQI), is an index the EPA uses to calculate and report daily air quality across the US and the associated health risks with Action Plans for those who are vulnerable. The AQI focuses on five major air pollutants regulated by the Clean Air Act:
- Ground Level Ozone*
- Particle Pollution (particulate matter)*
- Carbon Monoxide
- Sulfur Dioxide
- Nitrogen Dioxide
*Poses the greatest risk to human health
The AQI is calculated using EPA established national standards and is broken up into 6 categories using a “yardstick” that ranges in value from 0–500.
How Does the AQI Work?
Think of the AQI as a yardstick that runs from 0 to 500. The higher the AQI value, the greater the level of air pollution and the greater the health concern. For example, an AQI value of 50 represents good air quality with little potential to affect public health, while an AQI value over 300 represents hazardous air quality.
An AQI value of 100 generally corresponds to the national air quality standard for the pollutant, which is the level EPA has set to protect public health. AQI values below 100 are generally thought of as satisfactory. When AQI values are above 100, air quality is considered to be unhealthy-at first for certain sensitive groups of people, then for everyone as AQI values get higher.” https://airnow.gov/index.cfm?action=aqibasics.aqi
Important to Note: You can filter the values by zip code and the forecast is available daily and one day in advance. If the service focuses on air quality, we may need to reconsider the weekly updates feature that we considered previously.
Pollen and Mold
For pollen, we found a proprietary 12-point pollen scale from IMS Health’s Pollen.com which track pollen levels by day and offers 5 day forecasts by zip code. The scale is broken up into five levels of severity, from low to very high. This resource was particularly promising, as the forecast feature is aligned with our intended weekly forecast feature. Notably however, given the volatility of pollen levels in the environment, Pollen.com only offers notifications in two-day intervals when pollen levels are medium and above.
Finally, we found daily pollen and mold reports collected from dozens of pollen counting stations around the US by the National Asthma Bureau through the American Academy of Allergy, Asthma & Immunology. These reports contain data on levels of mold and pollen divided into four levels of severity, from low to very high.
The following sources were most useful for identifying data sources on air quality and pollen conditions:
- Air quality forecasts from AirNow developed to report local and national changes in the Air Quality Index (AQI) in partnership with The U.S. Environmental Protection Agency, National Oceanic and Atmospheric Administration, National Park Service, tribal, state, and local agencies
- Daily pollen and mold reports collected from dozens of pollen counting stations around the US by the National Asthma Bureau through the American Academy of Allergy, Asthma & Immunology
- Proprietary weekly pollen forecasts from IMS Health’s Pollen.com
- Hundreds of static datasets on air quality from a range of other federal and county sources at Data.gov
With these data sources in mind, we then mocked up spreadsheets to simulate the kind of data structures that our service might use, and how 3rd party environmental data might best be paired with user input fields with child-specific information about their triggers.
Alert Thresholds + Personalized Feedback
Because the AQI and NAB have different measurement scales and thresholds that trigger alerts, we discussed ways to distinguish how numbers on each scale would trigger an alert, and how personalized feedback would help the alerts become more personalized over time.
NAB — National Asthma Bureau (Pollen & Mold)
This scale is broken up into four categories (Low, Medium, High, Very High) based on particulate density per testing cite as compared to the national average.
We need to set parameters for what will trigger alerts within this scale.
AQI — Air Quality Index
The air quality index is broken up into 6 categories of measurement based on health risk. The thresholds between 101–150 put “sensitive groups” at risk, which includes children with asthma.
An alert will trigger when air quality is within this range. A feedback request will be triggered when the air quality is within the moderate range (51–100) to understand the sensitivity of the individual child. The uncertainty threshold will be within 50 points of the scale that will trigger a feedback request.
Simulating data structures
Using data from Pollen.com, we connected a Google Spreadsheet to Tableau and experimented with alternate ways of visualizing weekly thresholds:
- Size variation through squares
2. Size and color variation through circles
3. Horizontal and vertical line variations, with and without color variations
Our team found the horizontal and vertical bar charts with color variation to be the easiest to interpret.
Since we’re designing CoPilot as a mobile app, one concern is how to fit data for a week within the limited real estate of an app. Since it appears that most data sources show higher accuracy for the current or (at most) next day, we may dig more deeply into representing allergen and air quality data day by day.
After determining the exact data sources and time frame, the next step is to combine the sources and normalize the variety of scales into one severity scale for our service. We also need to examine the clearest way to provide fine-grained details about individual triggers.