Hi everyone,

I'm a fairly young Oracle + SQL Server DBA with DBA experience of
about 4yrs. I also posses OO programming skills in .NET and Java. As
you can see, I've been pretty much involved with both MS (Microsoft)
and non-MS camps, mind you, I'm an Oracle 9i certified professional
and lean more towards Oralce, Java, Linux etc.

Anyway, I've developed quite a fascination of late for the concepts of
data warehousing, data mining, knowledge management and I want to
pursue my career in this field as part of a larger pursuit of self
satisfaction.

Are you kind knowledgeable sorts able to guide me on how I can move
into knowledge management, without prior experience?

I was thinking of taking-up some sort of a Uni (or similar) course in
data warehousing and knowledge management and then applying for a
position in data warehousing (I'm assuming that I've got to have exp
in DW first before I could sink my teeth into data mining).

Also, I sometimes get disillusioned about the above plan in light of
the current deflated global IT market and think to myself that I may
be chasing an unrealistic goal, but then I read about how knowledge
Management, perhaps together with CRM, is of high importance to
businesses these days and will be even more in the future.

To say the least, I'm hell confused! So any help is much appreciated.

Many thanks
Toby

Data warehousing, Data Mining, and Knowledge Mgmt are HUGE fields.

IMHO,
Buy yourself a selection of DW fundamentals books from the old pro's:
Bill Inmon, CJ Date, Ralph Kimball, et. al. to get a broad perspective
of the topics to see what peaks your interest. If you are more
interested in back-end work spend some time understanding the joys of
missing data problem resolution, there is a very real need for people
who are committed to data quality (read some of Larry English's work
here) improvement methodologies and putting feedback processes in
place to improve operational data source integration with data
warehousing. Every day the data warehouse moves closer and closer to
becoming just as important an "operational system" to the day to day
business execs and field users as the transactional operational data
source itself.

I'd recommend you find yourself an opportunity to work with and
understand the different designs approaches necessary for ETL vs.
query performance. Perhaps as an assistant data architect for an
established data warehouse team to learn what it's like in those
trenches. Ultimately, designing for massively parallel performance is
still a bear, even with all the design tools available. Accomodating
the concepts of "slowly changing dimensions" in a data warehouse
design requires adherence to a disciplined design approach and
agreement with executive managment and project sponsors about just
what the scope of these items can be. Accomodating "slowly changing
dimensions" with a bolt-on solution can eat up an expensive ($10000
per GB) hardware budget in a hurry.

If however, you become drawn to the front-end or information delivery
end of things...
While the overall IT market has been depressed of late, Business
Intelligence opps have stayed pretty much stable. Spend some time
getting aquainted with OLAP and BI tools like Business Objects,
MicroStrategy, Microsoft OLAP Services or whatever it is they call it
now. These are all nothing more than information delivery tools, but
require someone who understands n-tier development, networking
thresholds, user-interface design, and can gather complete
requirements and translate them into finished, useful, purposed,
interactive products in the hands of the business folks in the field.

I ventured into the DW/BI field in 1996 and it's been going strong
ever since. Your OO skills will serve you well in requirements
validation and the ability to refactor a large solution into discrete
components.

RANT...
Java is still too damned slow for most high performance computing over
the network without native code integration, and dotNet is another
cheesy attempt by BillySoft to keep those of us who spent years
working with and around their technology from leaving the faithful. I
can only wish that one day EF Codd's SQL will go the way of all things
and be replaced by object applications with the ability to perform
whatever replaces relational calculus without the need of a database
engine...at least in the since/(sense?) that we know them today. XML
brings us closer to the generic, self-describing world, but at the
price of performance overhead. As more and more of the copper wires
turn to multiplexes of light todays performance problems will likely
be focused elsewhere. Nanotech semiconductors will ultimately allow
us to condense more and more storage into less and less space, which
corresponds to access speeds continuing to increase.

Good luck,
Don McMunn

toby_brown (AT) optusnet (DOT) com.au (Toby Brown) wrote in message news:<2a84f50c.0402180331.3147fbe8 (AT) posting (DOT) google.com>...

toby_brown (AT) optusnet (DOT) com.au (Toby Brown) wrote:
"I'm a fairly young Oracle + SQL Server DBA with DBA experience of
about 4yrs. I also posses OO programming skills in .NET and Java. As
you can see, I've been pretty much involved with both MS (Microsoft)
and non-MS camps, mind you, I'm an Oracle 9i certified professional
and lean more towards Oralce, Java, Linux etc.

Anyway, I've developed quite a fascination of late for the concepts of
data warehousing, data mining, knowledge management and I want to
pursue my career in this field as part of a larger pursuit of self
satisfaction."


Realize that data mining is fundamentally an analytical, statistical
process. Querying databases is a completely distinct function.

In data mining, one typically deals with data which is already
prepared as a single table (or at least abstractly, as a single
relational database query), and the goal is to have the computer
discover patterns in the data, as models, segments, etc. Input: "all"
of the data (or a statistical sample), output: the discovered
patterns.

Querying, on the other hand involves a dliberate specification of a
subset of the data to be retrieved. Input: query specification,
output: relevant data set.

While data mining may involve querying (especially to extract the
relevant statistical sample), and querying may be driven by things
discovered during data mining, these are seperate processes.

There is a large amount of high-quality material available for free on
the World Wide Web. I have found these search engines to be
especially useful:

AlltheWeb (www.alltheweb.com)
Dogpile (www.dogpile.com)
Google (www.google.com)
Teoma (www.teoma.com)
Yahoo! (www.yahoo.com)

Here are some searching tips:

Include "PDF" as a keyword: Documents in Adobe Acrobat format tend to
be better written and more polished than the usual stuff found online.

Include phrases such as "course notes", "class readings" or "lecture
notes"- these tend to find college professors' Web pages.

Also consider investigating KDnuggets (www.kdnuggets.com) and Citeseer
(www.citeseer.com), which is a search engine of technical papers.


Also, here are some titles I recommend:

"Computer Systems That Learn" by Weiss and Kulikowski
"Data Mining" by Witten and Frank
"Data Mining, Concepts and Techniques", by Han and Kamber
"Predictive Data Mining" by Weiss and Indurkhya

Books like these are reviewed in "Will's Technical Book List", which
you can find at:

http://will.dwinnell.com/will/willTe...lications.html


-Will Dwinnell
http://will.dwinnell.com

Thank you both D.McMunn and Will Dwinnell.

Your input will no doubt play a valuable role in my quest.

Cheers
Toby

predictr (AT) bellatlantic (DOT) net (Will Dwinnell) wrote in message news:<2b7b8021.0402200344.2faeed3e (AT) posting (DOT) google.com>...