Parcel clustering algorithm

gistfile1.txt

n = maximum number of parcels that can be in a cluster

Get list of parcels to survey. 
Create an empty list of clusters. 

Find top-leftmost parcel.
Find up to n-1 parcels that border that parcel. 
Only select parcels with the same street name. 
Only select parcels that aren't already in a cluster
Put the selected parcels in the same cluster.

If there are are less than n parcels in the cluster, repeat the process 
to find bordering parcels, but use the second parcel in the cluster.
(if no bordering parcels are found, what to do?)