Subject matter expertise addresses the degree of intimacy you have with the business rules and substantive meaning represented by the data you're processing. Visually, a text file of records may appear to just be lines of ASCII characters. However, knowing whether those records represent healthcare episodes, employment history, clinical trial phases, or financial transactions facilitates the ability to read the file contents correctly.
Knowing the story represented by the data helps you better assess whether observations, and the file as a whole, are consistent with the underlying business logic.
However, being a subject matter expert (SME) can have some disadvantages. A good reputation based on depth of knowledge in processing industry-specific data might get you to the top of the list when the organization needs an SME, but with it comes the risk of being pigeonholed. If you're known only as the person who can process specific classes of data, you may find your opportunities for assignments outside of your knowledge domain limited.
Product-line expertise addresses the breadth of analytics products you typically use to produce deliverables. Often, the type of environment -- tool rich or talent rich -- determines product line breadth.
A tool-rich environment is one in which the sponsors have chosen to invest in analytics products that will produce the desired results with a minimal amount of programming intervention. Organizations make this choice when the human capital pool is scarce or deemed too expensive to hire or train.
A talent-rich environment, on the other hand, is one in which the sponsors have chosen to invest in analytics professionals to produce the desired results, typically within a limited analytics product suite. Organizations make this choice when hiring or training programmers because it's more cost efficient than licensing additional products. It also happens when the desired deliverables are so complex or business-rule driven that customized programming is preferred over making a commercial product function beyond its intended purpose.
Platform expertise addresses the breadth of hardware architectures you use to do your work. The typical platforms are the mainframe, (also known as the Big Iron), midrange servers (such as Unix), and microcomputers.
While analytics software (such as is available from SAS, this site's sponsor) is consistent across platforms, each platform has components you have to consider to ensure that your applications function as expected.
Mainframes have a lot of processing power, but they tend to be expensive, so mainframe applications should be designed for maximum efficiency when execution time is a cost driver. Midrange servers are popular right now because they're less cost intensive than mainframes; they're also robust and fairly stable and reliable. The drawback is that using midrange servers also includes administrative and network security overhead; the native storage capacity is often less than with the mainframes as well.
Microcomputers have the least amount of processing power among the three, but they are physically portable and make data from personal and office productivity products analytics accessible.
The core issue concerning platform expertise is that you will have to develop some mastery of application tuning. Analytic products may execute on all three platforms, but the nuances and idiosyncrasies of each platform may impact outcomes such as execution speed, precision of default decimal values, processor utilization, and so on.
In effect, being an analytics professional means you have to decide whether to become a specialist or a generalist by subject matter, product line, or platform.
What choices have you made? What choices have been imposed upon you? Please share.