What does an area look like when you walk down the streets? Concreted and macadamised or leafy and tree-lined?
"Ancient trees...such things are beautiful; they have a high use which dollars and cents never represent. If the inhabitants of a town were wise, they would seek to preserve these things, for they educate far more than any hired teachers or preachers." (Henry David Thoreau)
​
The Green View Index ("GVI") described below is the first of three metrics that contribute to our aggregate "leafiness" score. The index itself is produced by MIT's Senseable City Lab as part of their Treepedia initiative; and what makes the GVI particularly valuable is that it captures the human, "street level", perception of leafiness. A summary of the GVI methodology can be found in the text that sits below the following map.
Please also note that this is our most data-intensive map (showing over 150k different points). Thus, depending on your speed of internet connection, it can take a couple of seconds for the map to reload each time you filter by a different postcode. Moreover, to truly get the most out of this map, you really need to view it on a laptop/desktop and not a mobile device.
The Green View Index ("GVI") is calculated using Google Street View panoramas. In the case of London, these panoramas consist of multiple street level camera images taken from 151,638 locations across the city during the summer months (May to September) in the eight years between 2008 and 2016. Each image is processed using the algorithm created by the Senseable City Lab. This algorithm calculates the amount of vegetation cover in each image and records this as a percentage figure. So, for example, if tree foliage covers one tenth of an image, the GVI for that image will be 10%. The images labeled a1 to d1 below illustrate this process.
​
It is important to note, however, that there are typically six different camera images for each location, with each one taken from a different angle of rotation. The GVI for any one location is, therefore, the average GVI of the six different images for that location, with additional adjustments made to take into account the angle of the camera tilt.
​
It is also worth observing that the GVI dataset doesn't map parks, as Google Street View doesn't venture into them as it does on typical city streets. Likewise, the GVI doesn't take into account the foliage in the back gardens of homes that are not visible from the street.
​
Although the computational burden involved sounds pretty daunting, the bulk of the heavy lifting had already been done by the MIT Senseable City Lab. Our job was simply to help display the data and (critically) incorporate the data into our aggregate "leafiness" score which you can use to filter London's districts in the Polymap and/or the Gallery.
​