This example demonstrates how to add emphasis to points on a plot/map to highlight clusters. The size of each circle is determined by a how much it shares a characteristic with its surrounding neighbours of a grid 3x3.

It uses colour encoding taken from from various examples of cloropethic maps based on d3's quantile scale. I've added a similar method for opacity.

The input is an array of objects. Each object must have at least 3 key:value pairs - 2 representing a x & y for your plot, the other should be a value of some kind. In this example the x/y is the easterly and northerly of 10km grid, the value is a rate between 0 and 1. An example:

[ {e: 37, n: 45, rate: 0.68},
  {e: 51, n: 49, rate: 0.48},
  {e: 39, n: 51, rate: 0.71}
  ...
  {e: 27, n: 64, rate: 0.62} ]

Some variables you may want to take note of:

• inverted sets whether the colour encoding is inverted or not, on a light background it should be set to false and conversely for a dark background. The example sets inverted = true;

• maxOpacity sets the maximum level of opacity for each point. The pattern sets the hottest regions as the least opaque. In this example maxOpacity = 0.90;

• stepsOpacity sets the steps for each opacity in the quantile output. This is how much less opaque one quantile is the its higher neighbour. In this example stepsOpacity = 0.07.

• scaleFactor sets the factor by how much the radius is scaled up in proportion to its similarity to its immediate neighbours. In this example scaleFactor = 1 - where all 8 neighbours are the same, the radius will be 8 x the default radius, a scaleFactor = 0.5; would 4 x larger and so on.