0

I believe there are three layers to build up- 1) Presentation Layer/Data Visualization Layer 2) Constucting olap server(like mondrian) to access mysql database and produce results. 3)Using Olap Client API(olap4j) to build OLAP cubes that store intermediate data.

Is this approach correct or is there any thing wrong with it?

Please suggest. Thanks in Advance

user961053
  • 1
  • 1

1 Answers1

0

Not quite.

  1. Build your data warehouse to Kimball standards. (Using ETL e.g. PDI/Kettle)
  2. Create your mondrian schema using schema workbench
  3. Get a BI Server up and running and publish your schema. You can then publish your schema and start doing analytics.
  4. Optional but recommended - Install Saiku.

If you can't do option 1 then Mondrian isnt for you. However you may be able to fake-up a kimball style schema with db views etc, but it wont work well for large amounts of data.

Codek
  • 5,114
  • 3
  • 24
  • 38