Tuesday, July 29, 2008

Slowly Moving Dimensions

Slowly changing Dimensions:
This is a small blog and its intended audience is everyone ( beginners and exp guys).By the topic of it ,I know readers must be surprised and say
“It’s a datawarehousing concept,so why is it here?”
You are right, it’s a datawarehousing concept and many of them ,who had worked on warehousing or who are familiar with data warehousing might be aware of this .
But , our question still remains
“What is it doing here?”
In our newer version of Essbase 11, of many new features , we have a new feature called “Varying Attributes”.This varying attributes is close to SCD’s of dimensional modelling ( SCD:Slowly changing dimensions).
The new feature of “Varying Attributes” helps a user in his analyses even further.I shall talk more on varyign attributes and how does it enhance the tool and helpful to user for his analyses in my next blog.
You can also find a podcast on “Varying attributes” in
http://analytiks.mypodcast.com/
Sandeep Reddy Enti

Tuesday, July 1, 2008

I am new to/learning Essbase, do I need to understand warehouse concepts? 1

Part 1:
The above subject is a question, I have been asked by few aspiring essbase learners. This article is for beginners. Expertise in data warehousing is not a pre requisite to work on essbase/ learn essbase , but conceptual clarity of the role of data warehouse in an enterprise is required.

Read and understand data warehouse to an extent, you can imagine where our OLAP engine fits in ,and more importantly to get a clear picture of the data flow. Here, I mean data flow from disparate data from heterogeneous data sources (which are in independent silos) to smart looking report, which provides valuable information and also instrumental in decision making .

The objective of this article is to make an image, which portrays the transformation of raw data into valuable information on reports.

Step 1: Let’s take an example of a bank. Bank has many departments, and each department has different application to process its departmental data. Where there is data, there is a database. As there are different departments, and different applications for specific needs, you have different data bases also.
Now, we are with different database (might be oracle, SQL server .DB2…etc).

Just imagine, few squares in different colors as few databases.
(I‘ll repeat, these databases have data from different departments of a bank)

Step 2: After drawing few colorful squares on left hand side , we have an ETL application right to it .ETL(Extraction , transformation and load). Its like a black engine, and has connecting thread to all the colorful squared (our databases) ones.
This ETL engine has the capability to pull the data, and transform into a common data type / desired format and load it into a common repository/storage place.
Eg: If I have 3 databases, and all these 3 database has information about a customer named ‘Srikanth’.
Database 1, might have stored as ‘Sreekanth’
Database 2, might have stored as ‘Shrikanth’
Database 3, might have stored as ‘Shreekant’
But its all of only one customer, so ETL will take care of all the cleaning required and put into storage, and make it sacrosanct.

So far, we have colorful squares (which are different database), right to it we have an ETL engine , which is strongly connected to the squares .Right to ETL engine ,we have a sacrosanct storage ,where the ETL puts in all the data.

Step 3: The repository/storage, which we had ended with in Step 2 is data warehouse.
Now, have an OLAP engine, I’ll fix it right to warehouse (ware house has data of all departments of bank).
Before I pull the data from warehouse and start building beautiful cubes. I’ll spend time on the final analyses and reports, which a decision maker/user would be interested to know of his business (here, its bank). When I am clear with the final report picture, I shall connect OLAP engine and pull the data, which is relevant to my final analyses/report purpose. Of the pulled data from warehouse, I build cubes.

Now, I have squared databases on the left hand side , an ETL machine connected to these databases .Right to ETL ,I have warehouse , the repository where ETL puts in data . Right to warehouse, we have our essbase engine (and cubes i.e. OLAP cubes).

Step 4: As users are not as smart as techies :), they need information in presentable manner. We build a reporting platform, and connect this reporting platform to our OLAP cubes. Reporting platform might be your Excel, analyzer, Interactive reporting….etc.
Now, we have a report platform, where you have information of the required analyses/report, which runs across the departments. This information has its geneses in the different databases.

The data has flown from different databases ,and ETL had put the data into warehouse ,and OLAP made even more beautiful cubes ,and the final reports, which a users looks at , is from the cubes.

Recapitulating ,
On left hand side, we have different databases, and ETL connects all, puts data into warehouse. OLAP talks to warehouse and makes cubes and provides information to reporting platform.

I could have simply pasted an architecture diagram, and made it simpler, but, now after this explanation, I think you guys wont forget the basic architecture and OLAP location in it.

As I had told you, its for beginners, If anyone is interested to know advanced topics of warehouse and role of essbase. mail me

One can find audio (MP3) on the same topic in the Podcast - http://analytiks.mypodcast.com/

Sandeep Reddy Enti