1.  Put together a block diagram for map reduce version of canopy clustering

2.  Put together algorithm to combine canopy clustering and k-means in order to make k-means faster. 

3.  Make suggestions about how to apply canopy clustering to EM gaussian mixture model.