Analytics Is Not The Problem
There is a great deal of attention these days on analytics, neural nets, deep learning and other advanced analysis methods. What is not mentioned is that these methods have been around for decades to centuries and they are only a minor part of the informatics problem. The access to contextually relevant data, that is normalized and consistent, is of vastly higher importance – as is the ability to deploy the results of the analysis of the data.
Complex analytics methods applied to data that is incomplete or inconsistent will provide poor results. Even simple analysis methods applied to broad, complete, consistent data will provide insightful results.
The second part of the equation – how to deploy the results of the analysis is also often neglected. It is important that the data science environment wherein the analysis is performed, must be able to create a standards-based, well-defined model that is portable and can be easily integrated into systemic solutions.
The problem to address is not analysis alone.