Did the reprieve from the November lockdown cancel Christmas?

23rd Dec 2020 dRISK

At dRISK.ai, we have been monitoring social distancing in real-time throughout the COVID-19 pandemic using CCTV feeds from London, and we have created a web-app available here. We had been tracking vehicle and pedestrian interactions for detecting transportation edge cases, and we repurposed this technology to detect social distancing across London at the start of the COVID-19 outbreak. Our tool measures patterns of mobility across London, and can identify social distancing bottlenecks and assess the impact of measures such as pavement widening on social distancing.

The announcement of the Tier 4 measures in London, in the words of some pundits, “cancelled Christmas.” This is incredibly disappointing for millions of families and those working in retail and hospitality. But increases in social mixing with the easing of lockdown in London correlate with increases in the incidence of confirmed cases of COVID-19: two consecutive weekends of shopping behaviour on Regent street as measured by our system are increase in confirmed COVID-19 cases seen in Government figures (Graph 1).

Graph of COVID-19
Graph 1 – shows rising rates of footfall in London’s busy Regent Street with corresponding increases in COVID-19 cases in London

There are a number of other contributing factors to the COVID-19 uptick than shopping behaviour on Regent street, such as the emergence of a new strain of the virus in the south east of England, thought to be up to 70% more transmissible. However, in general we have found that high levels of footfall indicate greater levels of social mixing and in turn less social distancing: we previously found a significant correlation between improved social distancing as measured with our London-wide camera-based system and decreases in the rate of new COVID-19 cases.

Busy Oxford St

The announcement of the Tier 4 measures in London, in the words of some pundits, “cancelled Christmas.” This is incredibly disappointing for millions of families and those working in retail and hospitality. But increases in social mixing with the easing of lockdown in London correlate with increases in the incidence of confirmed cases of COVID-19: two consecutive weekends of shopping behaviour on Regent street as measured by our system are increase in confirmed COVID-19 cases seen in Government figures (Graph 1).

There are a number of other contributing factors to the COVID-19 uptick than shopping behaviour on Regent street, such as the emergence of a new strain of the virus in the south east of England, thought to be up to 70% more transmissible. However, in general we have found that high levels of footfall indicate greater levels of social mixing and in turn less social distancing: we previously found a significant correlation between improved social distancing as measured with our London-wide camera-based system and decreases in the rate of new COVID-19 cases.

It is clear that social distancing will be with us for some time until the widespread rollout and uptake of effective vaccinations. During this time, big data techniques such as our approach in combining existing CCTV data and computation vision enables social distancing and patterns of mobility to be measured and analysed at scale. This provides feedback to businesses, government and the general public on levels of social distancing, enabling the economy to restart as safely as possible, ensuring a controlled exit from lockdown.

We will continue to develop these capabilities in the New Year, extending our tool with added functionality, including the capability to measure levels of face covering usage. We are interested in working with businesses and local authorities to deploy the tool during reopening in both indoor and outdoor settings, to provide a measurable, easy to use system to check if current measures to ensure social distancing are working as intended.