Visual Analytics Architecture

Hi all,

we are planning to build a new Visual Analytic tool. As we handle huge amounts of data, this is a very challenging assignment.
I am looking for best practices of architecture for the data layer supporting this new tool.
we have a very advanced OLAP cube, but I am afraid it will not be a good enough data source for the visual tool in terms of performance.
Our data is not sharded (in the OLAP cube).

Any best practices? we have a large DWH, OLAP cube and a Datamart. I am unsure if any of them can be a good enough datasource for a visual analytics tool.

Any help will be much appreciated.

Best practice is to look at what is out there and buy one... not develop one from scratch.

Why on earth would you try to build when you can buy? Surely Tableau or QlikView would suit your purposes?