Many countries have introduced social distancing measures to slow the spread of the COVID-19 pandemic.

To understand if these recommendations are effective, we need to assess how far they are being followed.

TheNewcastle University Urban Observatorywas established to better understand the dynamics of movement in a city.

How deep learning algorithms can be used to measure social distancing

It also monitors energy consumption, air quality, climate and many other variables.

These inform and update the dashboard in real time.

People Movement Monitoring Dashboard.

Article image

It’s free, every week, in your inbox.

In the graphic above, real-time data from pedestrian sensors is shown as solid lines.

The shaded areas are the normal pre-lockdown pedestrian flows.

Article image

Peaks in the graph represent an increased volume of people movement during rush hour.

Since the lockdown, however, only very small peaks have been observed overall.

Our research has found that pedestrian movement has reduced by 95% when compared to the annual average.

Article image

This shows that people have been following government guidelines closely.

This is possibly due to people shifting to using cars rather than public transport.

Overall, we estimate there have been 612,000lost journeys on public transportsince March 1 in Tyne and Wear.

Article image

This advice has been widely advertised, but it is difficult to assess whether it is being followed.

We produced models which can measure the distance between pedestrians in public places.

This was perhaps due to a lack of insight into widespread public behavior.

The Conversation

Observational infrastructure developed through technology may lie at the heart of future crisis management responses.

Smart cities use information and communication technologies to streamline urban operations on a large scale.

Citizens can engage with the smart city in a number of ways.

Data authority and governance will be an important point of discussion in future Smart City development.

Also tagged with