New model from Georgia Tech. What’s the Covid risk from a crowd? If you’re in a group of 50 you have a: 84% chance of being with someone infected if you’re in Atlanta 98% chance if in Phoenix 94% chance if in Austin Essentially 100% if in Charleston
Why avoid large in-person gatherings? This is a cool interactive tool from Georgia Tech: “The risk level is the estimated chance (0-100%) that at least 1 COVID-19 positive individual will be present at an event in a county, given the size of the event.”
After static releases of COVID-19 risk assessment charts, we are announcing an interactive dashboard that provides real-time estimates of the risk associated with gatherings of different sizes in light of ongoing (and often elevated) transmission:
Given the new case totals in the United States, if you have a gathering of 1,000 people or more, the odds are almost 100% that at least one person in attendance is infectious. Go to and see what your odds are based on different size events.
Today's update, the risk that one (or more) individuals are infected with #COVID19 in gatherings of different sizes. Map below for 50-person events, reinforcing the need to avoid large gatherings and take precautions (mask-wearing).
really, this is awesome! (or horrific, depending on your perspective) Durham County, NC, if i was to throw a 25 person birthday party tonight, 51% chance someone there has #COVID .... Miami-Dade? 10 person birthday = 74% chance....!?!😱
In Austin, there is a >99% chance that a group of just 100 people includes at least one #covid19-positive person. So is going to allow 50,000 people into the football stadium for games this fall. & h/t
Following up from this morning, some features of newly released #COVID19Risk interactive dashboard: State-level estimates of circulating cases (via APIs of ). Interactive selection of event size. Downloadable plots.
Major update to our dashboard, including real-time, county-level estimates that one (or more) individuals are infected with #COVID19 in events of different sizes:
Our dashboard has passed 1M visits. I'll have more to say next week, but for now, thank you to: , , , , T. Hilley (of ) + , , & .
Major update to , including two options for ascertainment bias: 10x (default, based on sero-survey medians) or 5x. Either way, event risk remains heterogeneous and particularly high in South, SE, SW, and W. USA:
Amazing risk assessment tool: I'm noting that Essex county Vermont, where my parents are, has among the lowest risk levels.
As of August 16, we estimate that Georgia has the highest per-capita circulating infection rate in the United States. That per-capita risk is the basis for risk-assessment estimates associated with gatherings.
This is a very cool model released by Georgia Tech. It shows the probability of #COVID19 being present based on the number of people at a given event. Basically, if you live in #Arizona there is a >99% estimated chance with events over 1k.
Also, thanks to and her new story in for telling me about this amazing project that shows you the percentage chance risk a person with COVID will be at a gathering in your county.
If you attend a gathering of just 10 people in Miami-Dade County the chances that someone there will have COVID is 79%
According to Georgia Tech's COVID-19 risk assessment model, there's a 63% chance that a gathering of 25 people in St. John's County would include at least 1 positive attendee... Group of 50? 86%. 100+ & it's basically a guarantee...
The odds of contacting coronavirus at any given gathering of people in the United States (in New York it's 40% in Miami it's over 99%)
High, heterogeneous #COVID19 risk in GA means local responses to #COVID19 needed institutions. Estimated risk 1+ infected in a class of 50: & (Fulton): 78% (Clarke): 73% (Bulloch): 90%
Starting today, you'll see risk maps generated at 5pm Eastern. These maps estimate the risk that at least one person at an event of 100 people is #COVID19 positive. Interactive maps are available at:
Rounding out the bottom 10, i.e., those with the highest estimate number of circulating #COVID19 cases as of August 18: 50. Georgia 49. Florida 48. Louisiana 47. Mississippi 46. Idaho 45. Tennessee 44. Nevada 43. Texas 42. Alabama 41. California
Our interactive dashboard takes real-time data from to provide state-level estimates of #COVID19 Event Risk Assessment: Take-away: masks, social distancing, avoiding large gatherings, can keep us all safer.
Major new release for : global expansion with 3 new country releases: UK, Italy, and Switzerland. Updated daily, estimates the chance that one (or more) individuals has #Covid19 in events of different sizes.
COVID-19 Event Risk Assessment Tool by ⁦⁩ from ⁦⁩. Now includes UK. Click on global risk estimates and then you can alter your group size and look for the risk of a person having covid at an event in your area.
For those waking up in a few hours on the other side of the Atlantic, our #Covid19 Event Risk Assessment Planner now includes France and Austria (+ UK, Switzerland, and Italy). Estimated risk that one (or more) have #Covid19 in events of different sizes.
Just shared and current risk of at least 1 #COVID19 positive person per 10, 25, or 50-person event with several dear friends in Lee County & Collier County, Florida (esp. in Immokalee, where I grew up). Got immediate feedback that...
We encourage folks with concerns to explore our county-level tracker (released July 7). Maps visualizes risk that one (or more) individuals has Covid-19 in groups of different sizes. Customizable and updated daily:
New interactive dashboard by w/ real-time estimates of risk at gatherings of different sizes factoring ongoing transmission. Addresses “what is the chance that 1 or more ppl at events of X size is infected?” Not instead of distancing/masks
Before reopening colleges and schools, check the probability that one of the students given the size of the group is #COVID19 positive. Currently, for Washtenaw County, MI: a class of 25 students: 21%; 50 students: 38%; 100 students: 62%.
So, let’s let data help with each situation... A free tool by at Georgia Tech can help estimate how likely you are to encounter someone with coronavirus, depending on the state you live in and how many people will be attending an event.
To get a better sense of the risk in those regions, I highly suggest visiting event-planning tool, which give the probability of a single infection coming to an event instead of the expected number and is updated daily 7/7
This is also a useful tool for calculating the risk of infection, based on a number of factors. (passed on to me by one of doctors I work with)
For those who missed this tool the first time, you can look up the odds of an event in your county having someone present with coronavirus based on the number of participants
Risk-estimates that one (or more) individuals in events of different sizes have #Covid19 in Italy... in Italian (): in English (): in mathematics: 1-(1-p)^n
There were over 1,500 (mostly maskless) attendees on the South Lawn last night. According to , in almost all of the US, the chance that no one out of 1,000 people is positive is less than 1%.
This is a great #covid tool (you can slide the event by # of people). It would be best if people understood a bit more about how probabilities work. Low #probability does not mean no #risk. (There was a low probability of Trump being elected, afterall).
For those interested in #COVID19 risk, we have released a UK-specific Event Risk Assessment tool: Presents chance that one (or more) individuals infected in events of different sizes.
For reference, in an event with "several hundreds" as the pool report in the linked thread says, group's Covid event risk estimator says 98%-99% of risk someone (other than Don) has COVID-19 in that crowd.
Someone has done basically that! It is at this website! (via/by )
A note re: and #Covid19. 8% positive surveillance suggests ~2.5K/30K students infected last week. If test capacity is 1%-2% of population per week, then reported numbers are likely *significant* under-estimates of true incidence.
Major update to #COVID19 event risk assessment planner, including daily state-level risk that one (or more) individuals is infected in events of different sizes, using 14-day data (and 5:1 undercount bias) as proxy for current incidence.
For folks tracking #Covid19 in the UK; we have extended our collaborative Covid19 Event Risk Assessment Planner to estimate risk that one (or more) individuals has Covid19 in gatherings of different sizes. See 'Global Risk Estimates' at:
We appreciate the significant and growing interest in the #COVID19 risk assessment calculator, and hope today's upgrade helps more folks access the site and take pro-active steps to reduce new chains of transmission.
Anyone from Mississippi aware of news reports on spatially clustered and elevated levels of incidence in W. MS? Image below given a group of 25 w/ascertainment bias of 5x (hence, a conservative choice, and suggests high levels of circulating cases):
With Fall term on the way, re-upping this story in which features our event risk tool: "Will I get coronavirus at this party?"
Via and in , featuring our #Covid19 risk estimates for UK gatherings: Story: Estimates (for UK and select EU countries):
With #Covid19 cases rising in the UK, you can directly access our global event risk estimates at: Maps show chance that one (or more) individuals has #Covid19 in events of different sizes.
Increasing risk in Midwest associated with large gatherings, thanks to and for story on collaborative Covid-19 Event Risk Assessment Planning Tool: Joint between &
A few pieces this week on school openings and risk. Want to highlight two themes that have been missed (and highlighted in our maps)
Given high #Covid19 case counts in the UK, want to explain how the 'rule of 6' is fraught. For more, see Thread below.
The most useful COVID-19 data breakdown: how likely are you to be exposed to SARS-CoV-2 in different cities, depending on size of gathering? HT
A reminder as GA expands its re-opening: the risk of infection remains elevated throughout the state. The vast majority remain immunologically naive, so wear a mask (esp. indoors) and distance when possible.
Thanks to for a mention of our event risk assessment tool on w/ last night. For more real-time updates of estimated risk that one (or more) are infected in groups of different sizes:
Risk assessment is a big piece of decision making. And this has been hard during COVID. With that in mind we've found this to be a helpful tool, FWIW.
Class starts in less than a month and by Georgia Tech's own modelling tool with the current numbers if all my students show up there's a 35% chance someone in the room will have COVID-19. How am I even typing this?
This is handy: #COVID19 Event Risk Assessment Planning Tool
COVID-19 Event Risk Assessment Planning Tool #covid19 #informatics
The risk: near certainty that one (or many more) in the crowd will have #Covid19. So, wear a mask and distance. And, if that's not possible -- ask yourself: is it worth the risk to you, your family/friends, and the community?
Summer of 2020: state-level variation in the estimated risk that one or more individuals had #Covid19 in groups of 50, using 5x and 10x ascertainment bias. Joint w/, , . Real-time, county-level updates at: