Re: Data mining tools

Jeff Swanberg wrote:

Hi Jeff,

1 - Given my example, do you have an opinion on the direction to go
(graphical query builder vs. cube)?

That's a bit hard for me to say. If you want to slice, dice, and drill
down large amounts of aggregate data, then an OLAP cube is a good way
to go. However, having re-read your original post, a query builder
might be best.

I actually worked with high school management systems for a number of
years, so I think I know where you are coming from. We were often
asked to create reports as you describe. Ex: List the children with
attendance below 85% that are living in single parent households and
have one or more siblings also attending the school.

While that report example _might_ be useful, it was incredible the
number of, IMO, completely silly reports we were asked to create.

Honestly, I've never put much stock in giving users ad-hoc reporting
capabilities, because one of two things usually happens:

A) You get people wasting incredible amounts of time (= money) trying
to create reports that the developer (you) could whip off in a fraction
of the time (= money). And often these reports are inaccurate because
people really don't understand the database structure or how to use the
tools at hand.

B) You get a "power user" that everyone else ends up relying on to make
reports. Not only is this person losing time from their actual job,
but if that person leaves, reporting capability is lost. Often, whoever
steps in to take over doesn't understand what the first guy did, so he
scraps everything the first guy did and starts over. This one happens a
lot with "Excel Wizards".

2 - I've never gone down the OLAP cube road. Do you have a
particular favorite where there are examples available and where the
learning curve to build the cube is not a vertical line?

I mostly use PivotCube because it's VCL for Delphi. When I started
working with PivotCube, the other VCL/Delphi products out there
(Express PivotGrid, RadarCube and FastCube) were not available at the
time.

I've had a cursory look at the other VCL offerings and none appear to
have the number of features as PivotCube.

The learning curve isn't vertical, but it might take a little bit of
time to wrap your head around it. The PivotCube docs aren't the best,
and a little is lost in the translation to English, but it's
manageable. Nobody ever accused me of being a "brainiac", and I got it
all working. :-)

Regards,

--
Kevin Powick
.

Relevant Pages

  • Re: How Large Is Too Large?
    ... That said it is not INXS of column counts allowed and if ... single Fact table with 200 columns realistic and feasible? ... Reports or via a cube or cubes? ...
    (microsoft.public.sqlserver.datawarehouse)
  • Re: Novice - Help with OLAP Design
    ... Designing time dimensions ahead of time is probably the standard way ... query or cube processing costs. ... your comment about knowing what the reports need to look like ... metrics they want to know as well, whcih is apart from report layout ...
    (microsoft.public.sqlserver.olap)
  • Re: Dynamically determine source measure in calculated field
    ... Server help. ... logic in the reports, as well as a ton of calculated members in the ... take that parameterization to the cube level? ...
    (microsoft.public.sqlserver.olap)
  • OWC export to excel
    ... I am displaying cube data in pivot table reports using owc in ... sharepoint services. ... names from owc by going to commands and options and changing the ... in the cube and ignores changes applied in owc. ...
    (microsoft.public.sqlserver.olap)
  • Re: SSAS reports, Proclarity and MOSS 2007
    ... we talk about cube and dimension security in SSAS himself. ... if you have multiple reports created in proclarity and you want to display only the list of authorized reports, then you have to apply a security to the report himself, or through MOSS, you can apply different security for each page / web site / webpart created. ...
    (microsoft.public.sqlserver.datawarehouse)