Eric:
Looking at Data Analysis within Wikipedia, a Data Analyst can do a variety data related tasks including Data Modeling to describe info, Data Mining, which includes data modeling for predictive purposes and Buisness Intelligence, which relates to aggregation of data (not sure what that is).
From what I have experienced, a Data Analyst is most often focused on the first, Data Modeling.
What is the relationship between Business Analysis and Data Modeling? Business Analysis is largely behavior (of a business system) modeling and data modeling. The key point is that behavior modeling drives effective data modeling. That is, a data modeling in itself is relatively inert: It is not going to guide the analyst through a rigorous data discovery. And before one can model data, they need to discover what all the data within scope is.
So, first, thorugh behavioral (process) analysis, we define our data flows, from which we create a dictionary of terms, (including data interrelationships), from which we can then do data modeling.
Those who talk about doing data analysis without first doing behavioral analysis (and there are alot of them) are assuming that data discovery has already been done, an all that is left is, essentially, to take the knowledge that is contained in a data dictionary and model it. This misleads as to the true nature of what hurdles are faced in creating a data model.
Tony