Old data, new analytics

Old data, new analytics

Professor George Roussos, Lead of the Pervasive Computing Group Birkbeck College University of London, blogs about the heat map they developed using data from the London Ambulance Service, which not only helps emergency operators visualise demand and coverage in real-time, but also predict growing future demand for their services.

Ambulances have different sensors that track their movements, like traditional motion sensors and each one holds multiple mobile interfaces with GPS. It was Marcus Poultons’, a PhD student at Birkbeck University, observation that there was so much ambulance data not being used, and he wanted to create a PhD project using their mobility analytics to find trends and patterns.

What we built was software that allows shift operators to visualise the movement of ambulances in real-time so that they can relocate the assets (ambulances) based on real-time coverage and demand. The data is presented as a heat map of London, which factors in real-time traffic data, the location of ambulances and their idle time. Predictions are made on a street-by-street basis and the colouring of the heat map is gradual, for example, yellow signifies where we expect first responders to arrive within 8 minutes with a small margin of delay, then the colours range all the way to red, which signifies where their ability to respond is well beyond the time limit.

heatmap

Screenshot of the software used by the London Ambulance Service

Monitoring and visualising London’s ambulance coverage in real-time enables operators to help first responders meet one of their main metrics, which is to respond to life-threatening events in up to eight minutes. Our software is also used by the two shift managers at the call centre so that they know when to give first responders a break and the apt time for them to go off their shift.

Prior to the introduction of our software, ambulance operators did not trust the automated tools available; believing that their suggestions and predictions were not accurate or adaptable to variables like seasons and demand during different days of the week.

Having Poulton work for the London Ambulance Service was one way he helped build trust among those using the service. He also ensured that the service was not disruptive or forced upon them to use. As a result of this approach, people began asking detailed questions about the software and Poulton was not only responsive but also candid about its limitations.

Growing demand

Like any other major city, the busiest day of the year for the London Ambulance Service is New Year’s Eve. In 2012 the number of people using the ambulance services on New Year’s Eve is now our current daily average in 2016. Within four years, what was exceptional has now become every day. These real-time mapping tools are critical for running a big city like London and while traditional services do not have particularly advanced data analytics, it is becoming more and more required.

Using historical data we can help the London Ambulance Service predict how demand will develop in the future, and it is something we are now building as part of our ongoing work to support them.

You can follow George Roussos on Twitter at @BirbeckUni. Don’t forget to follow IoTUK too @IoTUKNews

To read the full paper by Roussos and Poulton, ‘Towards Smarter Metropolitan Emergency Response’, click here

George Roussos
christiana.courtright60@cde.catapult.org.uk
No Comments

Post a Comment