[Mondrian] Lots of Dimensions & Lots of Rows

Paul Ramsey pramsey at refractions.net
Tue Feb 6 18:36:12 EST 2007


Hi All,

I recently stumbled onto the fact that the problem space I am working on 
is in fact a somewhat common and well understood one... the one where 
Mondrian is a solution. :)

The problem I am working on is providing a form of OLAP for multiple 
dimensions of spatial information in the province of BC.  For our 
design, each square hectare (a square 100m on a side) in the province is 
a fact entry.  That fact entry will have some measurements: length of 
roads, area of water, average slope, minimum winter temperature. It will 
also have some categorizations: what watershed is it in, what land use 
classification is it in, what forest administration area is it in.

It looks an awful lot like a standard OLAP problem.  The only thing that 
might make it non-standard that I can see is the data volumes and the 
query requirements:

Volumes:
- 110 million hectares in the province (so, 110 million facts)
- 100s of dimensions (as many as people provide)

Query requirements:
- What is the area of logging on steep slopes in the Blue creek watershed?
- What is the road density (length of road / (total area - water area)) 
in the Blue creek watershed?

My concern is that the number of dimensions on top of the number of 
facts will make the whole thing unworkable, since people are going to 
expect to query the thing through a *lot* of different paths.  As an 
added bonus, the deployment environment will be PostgreSQL.

Any thoughts from the experts?

Paul

-- 

   Paul Ramsey
   Refractions Research
   http://www.refractions.net
   pramsey at refractions.net
   Phone: 250-383-3022
   Cell: 250-885-0632



More information about the Mondrian mailing list