Analytic Method --> Level 1

Conceptually this section should focus around best practices for analytic project management in applied settings


Tools:


Training:


Web:
-------------------

I have been thinking about best approach to Method -- Level 1 sections for this blog. My goals are to point readers in the most optimal direction in their process of analytic discovery and in their process of discovering analytics. Today, we often learn that what industry tags as analytics is actually more appropriately named reporting, dash boarding, data mining, etc. When talking to decision makers, on the other hand, we find them unsatisfied with the analytic organization's output. Too often they are unable to manage outputs provided to do what they do best - make business decisions.

How do we distinguish components of analytic process and assign names to them?

Analysis does not even start with collection of data thru data mining techniques. Before we do that we should have specific business problem outlined and some hypothesis created to align project objectives with decision maker's needs. We will then figure out our data needs, decide on the appropriate analytic technique to solve the problem on hand. We will need to decide how to present the outcomes as well. When we combine all these activities we name it analytics because at the end it helps decision makers decide on "what to do on monday morning".

Here is the short list of things that Analytic discovery must contain in order to be considered analytics:

Analytic Project documentation - should include problem description, supporting data summaries (reports, dashboards, etc) and hypothesis around the stated problem. It is the quality of these questions that will ultimately the impact of analytic outcomes (how much better decision making becomes in competitive space). Reporting is an important step in analytic discovery process.

Data Mining and Data Strategy - Often called analytics by business intelligence vendors, it is a critical component of any analytic project and often a constrain in terms of final analytic method used, output created and quality of decision made.

Analytic Method - Once we know what the objectives are and data on hand we can make a decision on analytic method we will be using. There are often many methods that fit resolution of particular problem. The goal is only one, achieve repeatable outcome with stated degree of accuracy.

Presentation - Likelihood that three steps above will provide actionable insight is still pretty low. Your goal is to present the insight in concise format so that decision can be made quickly and confidently (on that same Monday morning).