Like light from a distant star that takes years to reach our eyes, data from the NHS is a snapshot of infections that happened 10+ days ago. To know what's happening *right now*, this site uses an interactive model to estimate what happened during that time.

Inner London

0

0

Greater London

0

0

England

0

0

Date:

Data from Public Health England. Shows situation as of at 0900. New data released every evening.

This model is editable. Drag the blue numbers to update the estimates!

Londoners are catching a disease called COVID-19. As they were getting on with their lives, they were sadly infected with the SARS-CoV-2 virus.
days later they began to notice their first symptoms.
After days of being ill, % of them were able to get tested. They got the results day later.
This means that the cases that the NHS has confirmed in London as of morning were first infected around days ago.
Unfortunately, during that time the virus has been spreading exponentially, at an estimated rate of % each day. People are highly contagious for days after symptoms appear...and day before.
With these assumptions we can estimate that there are ** infected people** in London right now.
** people** don't have symptoms yet.
** people** are contagious.

The estimates are generated using a simple model of exponential growth. The goal of the model is intelligibility, allowing users to explore the data for themselves. The key assumptions of the model are:

- The NHS data captures infections that occurred many days ago. This delay between infection and data being reported to the public is a constant.
- The number of cases reported by the NHS consistently underestimates the actual number of cases by a scale factor. This is phrased in the model as the % of people who are sick who get a test.
- The virus is growing exponentially at a consistent growth rate. Each region grows at this rate. Local estimates are not used due to noisy measurements.

The rough algorithm for finding the number of infections today is as follows:

- Find
`scaledConfirmedCases`: Scale the latest number of confirmed cases by an estimate of what % of people have been tested. For example, if the NHS reports 1965 cases, and 50% are tested, scaledConfirmedCases = 3930 cases. - Find
`delayDays`: Estimate the delay between infection and data reporting. For example, if incubation period is 5 days, people get tested after 2 days, and it takes 1 day to get results, then delayDays = 8. - Find
`growthRate`: Calculate an average daily growth rate for London. This can be estimated by looking at the growth in official count of cases, or alternatively, by the growth in the official death rate. - infectionsToday = scaledConfirmedCases * growthRate ^ delayDays. For example, infectionsToday = 3930 * (1.27)^8 = 26596.

How Nearby Estimates Work: The number of nearby cases is calculated in two steps. First, the number of infections in each borough is estimated using the exponential growth model described above. Second, the number of cases nearby is calculated using a simple Monte Carlo method. The postcode is converted into a latitude and longitude point. A circle is drawn of radius 400m (for the 5 minute walk) and 1200m (for the 15 minute walk). For each borough these circles touch, a set of random points equal to the number of estimated cases is generated. The number of cases inside the circles are counted. This is done 1000 times. The average represents the number of cases nearby if each case is spatially isolated. Evidence from other cities shows cases of COVID-19 occur in family/household clusters. This pure average is divided by 3 to generate a range, which reflects the situation if the number of cases per household varies from 1 to 3.

Borough | New | Total |
---|