Nearly 250,000 people left London before the 23rd March Lockdown

There has been much speculation over the possibility that large numbers of people left major cities in England, particularly London, for other areas of the country in the weeks before the government’s instruction for the British public to stay at home and socially distance, announced on 23rd March. Anonymised and aggregated GDPR-compliant location data drawn from opted-in mobile phone users shows that this trend did occur in both London and Birmingham but not to any great extent in Manchester.

On 16th March, the day that the government asked households to self-isolate should any single member show symptoms of coronavirus, just over 100,000 people had left already London for other parts of the country. This figure rose to just under 150,000 on 18th March, which was the day that school closures were announced to begin in two day’s time. When schools closed across the UK at end of the day on 20th March, total net outflow from London had risen again to stand at just over 200,0000 and three days later, when the Prime Minister asked people to stay home for all but essential reasons, the total reached 237,000. Since then, the net outflow from London has crept upwards over later March and into the first half of April, standing at 262,000 by 12th April. This equates to an additional 25,000 people leaving London for other areas of the country since 23rd March government instruction to socially distance, showing some subsequent movement post-lockdown.

Even so, these overall pattern shows that the vast majority, or 90%, of those who have left London since March did so prior to the government’s instruction to reduce movement for all but essential medical, food shopping, and work-related reasons. Since the data is completely anonymous and aggregated, it does not provide any indication of the reasons why people chose to leave London or why those chose to travel to the destinations that they chose. However, it may be reasonably speculated that people may have done this for any number of reasons, including the possibility of reuniting with family and loved ones located in other areas of the country in anticipation of the spread of the national coronavirus epidemic and the government’s increasing nudge towards social distancing prior to 23rd March.

Notably the evidence suggests that people representing only a very smaller proportion of London’s total population, about 3%, have left London since March.

The British public has been increasingly following social distancing

Adherence to social distancing has been on the rise since the Prime Minister's 23rd March announcment for the British public to 'stay at home' in an effort to contain the spread of coronavirus. On that day the percentage of Britons not moving was just under 40%. This share of 'non-movers' has significantly increased over the following weeks, reaching a high point of 60% on Easter Sunday. On Easter Monday, the figure had dipped slightly to 55%.

The data shows that the government's phased introduction of social distancing measures were successful
‍
The British public were responsive to social distancing as new measures were phased.

On 16th March, the Prime Minister announced two instructions to members of the public. Firstly, all households in which any single member was displaying coronavirus symptoms were asked to self-isolate and to stay at home for two weeks. Secondly, non-essential contact with others and all unnecessary travel was suspended.
‍
With this announcement alone, UK population movement sank by 21 percent. With the announcement of school closures just two days later, population movement dropped by a further 37 percent in the following days.

As the chart below shows, the lion's share of population movement decrease had already taken place by the Prime Minister's announcement on 23rd March of a full scale national lockdown. However, this instruction saw an additional fall in people's movement by another 26 percent to its current level of 87 percent below the pre-crisis movement patterns. Based on the data, it would appear that the British public had already been nudged in this direction before the 23rd March lockdown instruction by the preceding self-isolation and school closure announcements likely combined with media coverage on the global pandemic.

Population movement stayed low during the extended Easter weekend
‍
Compared with the previous two Saturdays, Easter Saturday exhibited a slight increase in population movement. Note though that the colour bar has a maximum of 0.4 kilometres, which is well under the usual population movement in excess of 1 kilometre.

Less pronounced drop in population flows compared with population movement

While the rail journeys in the UK are now at just five percent of normal levels, population flows between regions only dropped in its peak around 80 percent. This suggests a shift in the mode of transport from public options, such as rail, bus and tube to private means of transport by car. The drop in population flow is also notably lower than the drop in population movement.
‍
We measure population flows as the percentage change in the people entering the destination clinical commissioning group from a different origin area. As the population practices social distancing, we expect to see fewer border crossings.

Supermarket footfall shifts from evenings to mornings

Supermarket footfall across the United Kingdom exhibited minor shifts from the evening hours to the before noon hours. With usual shopping patterns, the lowest footfall in a supermarket is 9 am on a Sunday. However, during the lockdown, the footfall on Sunday mornings has substantially increased. For the last week, the best time to go grocery shopping is Tuesday mornings from 9 to 10 am.

Despite surge in ICU demand, overall drop in hospital footfall

Visits to Public Parks have decreased less than overall movement

Risk rank based on age, respiratory disease and cardiovascular conditions

The risk rank is determined by two components: 1) percentage of individuals aged 70 and above, and 2) percentage of individuals aged 16-69 and below with any of the health conditions in respiratory diseases (such as asthma and chronic obstructive pulmonary disease), cardiovascular diseases (such as stroke, heart attack, and angina, diabetes and hypertension).

We expect the peak of infections for the UK in the first week of April

Forecast of the number of daily new confirmed COVID-19 cases for three regions: UK as a whole, London and aggregate regions outside London.  NHS actual daily confirmed cases are shown as ‘+’. Median, upper and lower quantiles of predicted cases are shown as black, red and green lines respectively. The model is a simple non-stochastic compartmental SEIR. Confirmed cases are assumed to be proportional to the exposure (i.e. the E in the SEIR model), subject to some randomness. We are not epidemic modellers and have no training in epidemiology.  Our forecasts, therefore, should be treated with a high degree of scepticism.