Ask HN: What specific tech. are there to perform cluster analysis wi/ bitstrings

Let's have a use case: climate classification.

I have a bitstring of many (50 bits long? 500 bits long?) binary classifications (e.g. does it regularly hit below freezing here during winter? Is there permafrost here?, etc...) and I have a million points sampled, resulting in a database with a million 500-bit bitstrings.

Now, if a geographic point has 1 on the "is there permafrost here?" slot, then it's probably also going to have 1 on the "does it regularly hit below freezing" slot. So clusters are going to form.

Now there has to be a way to group these points into partitions, and I need a criterion for "the best kind" of partition, tailored for bitstrings. How should I go sbout with this?