Alex Rich and Cameron Yick built this as an all-volunteer project.
Alex is a PhD candidate at UNC Chapel Hill’s Carolina Health Informatics Program (CHIP). He received a Masters in Public Health from Yale where he was a teaching fellow in healthcare operations and a healthcare operations innovation fellow at the Yale Center for Biomedical Innovation & Technology (CBIT). Prior to that, he was a Major, Senior Pilot, and crash investigator in the US Air Force Special Operations Command.
Cameron is a full stack developer and data visualization specialist at Datadog in New York City. He received a BS in Computer Science and Electrical Engineering from Yale. He is regular contributor to open-source projects, especially when they relate to interactive graphics, making public datasets more accessible, or health and environmental causes.
The team’s first collaboration was as part of Team OPAT, winners at the US Dept of Health and Human Services Data-for-Opioids Code-a-thon.
Both Alex and Cameron have younger sisters working in hospitals on the front lines of the Covid-19 pandemic.
We built the first version of this tool after politicians around the country started talking about sacrificing American lives for the economy by sending people back to work. Initially it was an effort to watch over and communicate with friends in Texas. We then recognized that the need was national and scaled the tool.
We want people to be able to see the threat around them, personalized to them, and make a commitment to taking the steps they can to slow the surge of patients threatening to overwhelm hospitals around the country.
We both have little sisters working as caregivers on the frontlines.
Alex used to work as a special operations pilot and crash investigator in the Air Force. In every crash or mishap he investigated, there was one moment when the right piece of actionable information could have prevented disaster.
We thought about what that piece of information is in this pandemic and decided that the most important information for behavior change is a sense that the threat is real and local.
We see a lot of bubble maps out there, but nothing that takes away the mental math of trying to count up and compare and estimate what’s around the user.
We built that.
Next, we conferred with a behavioral economist who recommended a more specific and actionable “ask” and commitment device... a tool to help lock in the desired actions in the user’s mind. Commitment devices help close the gap between what a person intends to do and what they actually do.
We built one based on CDC recommendations, and added a social component by letting users share their Covid Commitments on Facebook and Twitter.
We start by pulling daily updated confirmed case and death counts from the Johns Hopkins CSSE Covid-19 Repository. We merge this file with a shapefile of US counties by FIPS code, and WGS84 Country Shapefiles. Note that the US is the only country where we currently support subnational regions.
Next, we use the Algolia Places and Google Maps APIs to convert the user’s entered location to coordinates, and to determine the state/country of that location.
Because the Hopkins data combines several counties for New York City, we calculate if the user’s entered address is within 100 km of New York City. If so, we merge the shapefiles for those counties into a single polygon.
Afterwards, if the address is in the US. we use the OpenRouteService API to generate a 1-hour drive time isochrone around the address. We use MapBox’s API as a backup in case of rate limiting or capacity issues. We collect all counties that intersect with the 1-hour drive time isochrone polygon.
We tally the cases and deaths in all of the intersected counties for display to the user as tooltips, as well as an aggregation for the state that the user’s entered address falls in. We visualize the counties and isochrone using the Python Folium library in a
We use this same data from Johns Hopkins to calculate the national totals for cases and deaths. This total includes the rows of data that lack FIPS codes or county names.
Finally, we export a
Leaflet.js map and several
JSON file with the 1-hour-drive-time, state, and national confirmed case and death totals, which are sent to the user via a
Next.js React web application.
The team is still working to fine tune the tool and to get as many people as possible to make the Covid Commitment.
We are actively seeking advice on how to better engagem with both social media and traditional media. Please email
email@example.com to get involved.