Fires of London: Study 1

temp1_watersmall

In short: Each circle is one incident; the whiter the color the quicker the response time to that exact position.

Description: The average responder times for the London Fire Brigade. The darker the color the longer the time, the hue shows the difference between the arrival times for first and second fire engines. Where New York might be synonymous with the yellow cab, London is inevitably associated with the Great Fire of London of 1666. This image has larger overlapping points to give an idea of the general response times, whilst maintaining the scale and features of the London landscape. As a result it is clear where the locations of the fire stations in London are, despite this being absent in the data itself. Having small discrete points in this case is meaningless as the response time will likely be similar to the next street along and (fortunately) the incidents are not so numerous to gain any information on a smaller scale. The area-of-effect of the data is key here. The resulting image shows much more than discrete data points, and fixing the red channel gives it a clear impression of fire, as though we are looking at the glowing embers of London.

Point of interest:  The slightly illogical aspect of this image is that the white (hot) is near a fire station, whereas the orange/red (cooler) is further away – when it should really be the opposite. This may seem wrong and was chosen to give the image some identity. In reality though, a good proportion of fire brigade responses are not actual fires, including amoungst other things, false alarms in communal residences from students (drunken status: data not available) – 4801; monasteries/convents – 44. And animal rescues, including one instance of rescuing a kitten whose head was stuck in some bongo drums (response time: 3 mins 39 seconds).

Technical: All 485,056 incidents responded to by the London Fire Brigade 2009-2012 (data available at https://data.london.gov.uk/dataset/london-fire-brigade-incident-records). Each incident was added geographically to the image with an equivalent point-spread of ~0.1 miles with the average response time calculated for each point. These geographic response times were combined for both the first responding engine and the second if one was required. The red channel was fixed, with the green and blue channels representing an average of the first responder time and a component of the first responder time and second responder time in the green and blue channels respectively (normalized to the maximum time a second responder arrives in). In effect the darker the orange/red, the longer the time for the engines to arrive, if there is a pink hue then the first responder arrives at a similar response time that would be expected of a typical second responder. Yellow is where the first responder arrives promptly but experienced a delay on the second responder.

Comments are closed.