Help Predict COVID-19’s Spread in Your Community
Leading epidemiologists are assisting the CDC and other federal and state agencies to develop computational models of how COVID-19 might spread in your community. These models depend on knowing the different social distancing measures in each of the 3000+ counties in the United States. We urgently need volunteers to help us build a dataset of social distancing measures.
Progress so far
We currently have data for US counties. Counties in green have complete info, counties in yellow need answers from local residents, and uncolored counties need your help! Click on a state to find data for your own county!
Due to a high rate of response from volunteers, we are taking longer than usual to verify and post data to our site. Thank you for your contributions! Please continue to gather information and share the site - we still need your help to reach our goals! Counties with unverified data are shown in pink.
We are in the process of integrating data aggregated by other researchers and volunteer efforts as well. So far we have integrated data on K-12 school closures, shelter-in-place measures and business closures for 158 counties in Georgia, gathered by the University of Georgia CEID COVID-19 Working Group.
Add social distancing information for your community
Help us build better models of COVID-19 by providing data on social distancing across the country in the survey below!
You can start by adding data for your own county. If you want to help even more, you can look up info for other counties online.
FAQ
How is this data collected and verified?
All the data on this site has been collected from public news sources and government communications, either by our team or by volunteers like you. We manually check all the data contributed by volunteers.
I see incorrect or out-of-date information. What should I do?
The situation on the ground is changing rapidly. We expect that the data will need to be updated regularly. Whenever you see something that needs to be updated, please let us know using this form.
What happens after I submit my contribution?
Our team is on call several times per day to verify the information that is submitted. After verification, we publish your contribution live on the website and in the public dataset. We appreciate your contributions!
What if different places of worship, colleges, universities or K-12 schools in a county closed on different dates?
Counties may contain multiple colleges, universities and school boards that put in place different rules along different timelines. Similarly, religious gatherings have often ceased voluntarily, in which case different congregations may have different timelines. In this case, we ask that volunteers report the earliest closure date they are aware of in that county.
What if there are different rules (such as business closures, stay-at-home or shelter-in-place orders) in effect in different parts of a single county?
Counties often contain one or more municipalities that may put in place their own rules that are stricter or different than those that apply at the county level. In this case, we ask that volunteers report the most restrictive policy that is in effect in any part of the county.
How did you choose which data to collect?
We collaborated with epidemiologists producing models of local spread for COVID-19. The items being collected, including shelter-in-place orders, local religious gatherings, school closures, business closures, and the start dates of each of these, are the central components of their models for predicting COVID-19 spread. We opted to keep the data collection as lean as possible, the biggest "bang for the epidemiologists' buck", so that we could successfully collect across as many counties as possible. This information is so important because it enables more accurate model fitting and results in better estimates of COVID-19 infection rates and how much they are reduced by each policy.
How is this project different from the many other COVID-19 projects?
We are aware of parallel efforts, including those from Keystone and Wikipedia, who are also gathering county-level data on nonpharmaceutical interventions. We communicated with some of these efforts to synchronize our data schemas, and we are committed to sharing our verified data publicly as quickly as possible. Integrating data from other sources into our crowdsourcing workflow to enable data verification is one of our immediate priorities, and we welcome others to use the data we produce as well. Everyone's goal here is to get to a complete and accurate dataset as quickly as possible, and we welcome any efforts that allow us to work together toward that goal!
Why is crowdsourcing a good way to collect this data?
There are over 3,000 counties in the United States. Wikipedia, Keystone, and others have on the order of tens to a few hundred counties covered. We need much, much more data in order to succeed at this epidemiological task. Luckily, this is also data that is known to residents of each county and searchable via the web - which is why we need your help!
Can we use this data set? Is it licensed?
All the data is released under a CC-BY License, meaning that it can be used free of charge for commercial and non-commercial purposes. However, under the terms of the license you must acknowledge our project as the source of the data. Please let us know if you use our data in your project, we'd love to see what you build!
Who is behind this project?
We are a team of computer scientists and epidemiologists who are assisting the CDC and other federal and state officials in their COVID-19 planning and response efforts, and our work is briefed to senior federal officials. The epidemiologists on our team have previously supported responses to the 2009 flu pandemic and other epidemiological crises. Funding for this project was provided by a National Science Foundation Expeditions Grant (Sub-Contract funding (ARA): Scalable Analytics for Decision Support), and by Amazon Mechanical Turk.
Our team: Jacob Ritchie, Tum Chaturapruek, Mark Whiting, J.D. Zamfirescu-Pereira, Mitchell Gordon, Jackie Yang, Tianshi Li, Amy Zhang, Catherine Mullings, Rajan Vaish, Golrokh Emami, Danae Metaxa, Mandy Wilson, Achla Marathe, Stephen Eubank, Madhav Marathe, and Michael Bernstein.
You can contact us at covid-crowd-contact@lists.stanford.edu.