Chapter 14. Inside Analysis Services Data Mining Tools
Microsoft SQL Server 2000 provided one main tool for developing and managing mining models. Analysis Manager was the main tool. Built on the Microsoft Management Console (MMC) infrastructure, it enabled the database professional to develop and manage mining models against a live server. In this tool, design and management tasks were combined in one unique environment.
With SQL Server 2005, Microsoft has redesigned its tool approach to better fit different roles played by different individuals or teams in a company. Most functionality is now split between either the Business Intelligence Development Studio for design tasks and SQL Management Studio for management and other operational tasks. When we started early designs and prototypes in 2000, the decision had not yet been made to integrate the BI development environment with VS, but very early it was clear that we wanted to provide an experience to the "BI developer." Unlike the developers using the rest of the BI stack, data mining developers often are analysts, rather than real developers. So the choice to integrate with Visual Studio may sound odd, but it was made so that all the tools are consistent across the stack. Although the analyst needs to work in a Visual Studio shell, he or she will not at any time be required to do any coding.
This is why every single task can either be achieved through graphical wizards and designers, as well as through scripting and coding with the APIs. We do realize that even though we have made tons of investments in providing very business-oriented wizards, the IDE environment can still be somewhat overwhelming for the non-developer. Look at the development tools in SQL 2005 as the first step into this direction. I'm sure that as we learn more from customers' experiences and hear more feedback from users and developers, we will refine and provide tools in the future that are even better suited to individual roles, levels of expertise, and tasks.
This chapter begins with an overview of Analysis Services Tools and then tries to walk the reader through many scenarios that cover most of the key functionality of the tools while also providing insights, tips, and techniques. I hope readers will find a lot of valuable how-to techniques throughout this chapter.