The square based clustering method left something to be desired as the square were quite obvoius, so I tried a new tact. Also it isn't a good idea to test clustering on random data so this is the boston police log.
Modified http://d3js.org/d3.v3.js to a secure url
Modified http://cdn.leafletjs.com/leaflet-0.5.1/leaflet.js to a secure url
Updated missing url https://raw.github.com/mbostock/d3/5348d911938a0d1fdf43d7c86befbd908e431204/lib/colorbrewer/colorbrewer.js to https://cdn.jsdelivr.net/gh/mbostock/d3/5348d911938a0d1fdf43d7c86befbd908e431204/lib/colorbrewer/colorbrewer.js
Updated missing url https://raw.github.com/Caligatio/jsSHA/release-1.42/src/sha1.js to https://cdn.jsdelivr.net/gh/caligatio/jssha/release-1.42/src/sha1.js
https://d3js.org/d3.v3.js
https://cdn.leafletjs.com/leaflet-0.5.1/leaflet.js
https://raw.github.com/mbostock/d3/5348d911938a0d1fdf43d7c86befbd908e431204/lib/colorbrewer/colorbrewer.js
https://raw.github.com/Caligatio/jsSHA/release-1.42/src/sha1.js