All about using Microsoft products to deliver end to end BI : Analysis Services (SSAS)

The Microsoft Data Warehouse Toolkit.. Theory meets practice

raheelr — Sat, 25 Mar 2006 21:04:00 GMT

I had been waiting for Amazon to send my copy of Joy Mundy's, Warren Thorntwaite and Ralph Kimball's new book and let me tell you after going through it for the past few days, I am extremely excited of the potential all customers that I work with will have as they work on their own DW/BI system initiatives.

If you ever wondered after reading a theory book how am I ever going to apply this theory to the tools that I have to work with? Well, this book answers that exact question as it takes the best of the conceptual insights from Ralph Kimball's Data Warehouse Life Cycle toolkit and walks through how to use those insights with all the components of SQL Server 2005 including ETL practice with SSIS, data modeling, OLAP, APPLIED data mining, reporting and BI applications.

Here is a link to save you the trouble to finding it on Amazon :) http://www.amazon.com/gp/product/0471267155/sr=8-1/qid=1143313605/ref=pd_bbs_1/103-4596045-3657420?%5Fencoding=UTF8

-Raheel

Mystery of the dissappearing roles in Analysis Services 2005..

raheelr — Sat, 25 Mar 2006 20:54:00 GMT

One of the major differences in working with Analysis Services 2005 as compared to Analysis Services 2000 is that there is a specific developer environment to develop your cubes, data mining models etc and then there is a seperate interface for managing your Analysis Services 2005 server; SQL Server Management Studio.

This is different compared to AS 2000 as there was just one tool for developing and managing your cubes Analysis Manager. Therefore is some ways if you are used to the older way of doing things you would create your security roles using Analysis Manager. But with seperate interfaces with SSAS 2005 which tool do you use for that purpose?

Well, I have been fielding questions on dissappearing security roles within your SSAS environment. If you used the BI Dev Studio to create your cube and deployed it to the server and then used the Management Studio to assign security but then you went back to the BI Dev Studio and rebuilt your project; guess what everything on the server is rebuilt and therefore it overwrites your security assignments. AHA..

You either assign security roles withing your BI Dev Studio package so if you ever rebuilt your cube the security structure is rebuilt or you dont always rebuild the structure of the cibe just the data ensuring that the security setting always stay on the server..

-Raheel

Excel 2007 and Analysis Services 2005

raheelr — Sat, 25 Mar 2006 20:50:00 GMT

I just installed beta version of Excel 2007 and there is a lot of interesting integration with Analysis Services 2005. For a list of enhancements to Pivot Tables check out the Excel Blog at http://blogs.msdn.com/excel/archive/category/11544.aspx. I will have additional posts as I learn more.

Building Report Models on top of Analysis Services

raheelr — Tue, 03 Jan 2006 16:47:00 GMT

There is a new capability within Reporting Services that allows ad-hoc reporting for end users directly. That tool is called Report Builder and can be invoked through the Report Manager web interface. In order to deliver this capability though to the end user first thing you need to do is provide the user with a Report Model. A Report Model is a very intelligent semantic model that can currently be built on top of multiple SQL Server or Analysis Services sources. The goal of the Report Model is to provide a subject area definition of the data that a user has access to along with friendly names and description of data for easier Report Creation.

The Report Model is a specific type of a project available in the Business Intelligence Development Studio ( BI Debv Studio), the development environment shipped with SQL Server 2005.

A lot of customers are excited about this capability and rightly so. Report Builder is a fantastic tool and can add a lot of value to the overall BI stratgey of an organization. The question people usually ask me is how do I build a model on top of Analysis Services.

Well, if you try and create a new Report Model in the BI Dev Studio, you will notice that it only allows you to create a model on top of a SQL Server source. It does not allow you to create one on top of Analysis Services. The primary reason is that an Analysis Services cube already has a lot of the model capability built in it, like understanding of hierarchies, attributes etc that the model can leverage. Therefore in order to create a Report Model on top of Analysis Services this is what we need to do:

1) Open the SQL Server Management Studio. This is the management tool inlcuded with SQL Server 2005 that manages all the components of SQL Server. This includes the relational database, Analysis Services, Reporting Services, Integration Services and SQL Server Mobile databases. Register your Report Server in the Management Studio. TIP: To register your Report Server with the Management studio provide name like : http://<servername>/reportserver

2) Once registered you will be able to traverse the objects in Reporting Services like your Data Sources folder, ReportS folder etc. Rigtt click on the Data Sources Folder and make a new connection to your Analysis Services cube.

3) Once the connection is made, right click on the connection and one of the options you will now see is to Generate Model. Thats it.

Once the model is generated a user can use Report Builder to use all the rich hierarchies and attribute availability to build their own reports.

Managing Metadata : SQL Server 2005 Meta Data Samples Toolkit

raheelr — Sun, 01 Jan 2006 22:59:00 GMT

Meta Data is an area where a lot of customers are looking for a solution. SQL Server 2005 components specifically Integration Services, Analysis Services and Report Model capabilities in Reporting Services provide a mechanism to bridge multiple backend sources for either data integration purposes, analysis purposes or reporting purposes. The question comes when a need to understand data lineage or dependencies between the data is required. In this case, this recently posted Metadata Samples Toolkit aims to provide that information to the BI team.

I have not had a chance to look into it in detail yet. I do plan to very soon and will share the details as soon as I do.

http://www.microsoft.com/downloads/details.aspx?FamilyID=11DAA4D1-196D-4F2A-B18F-891579C364F4&displaylang=en

Happy New Year everyone!!

Raheel