Scope of development
Data analytics workflow can be roughly divided into two phases, the build phase and the operationalization phase. The first phase is usually a one-time exercise, with heavy human intervention. Once we've attained reasonable end results, we are ready to operationalize the product. The second phase starts with the models generated in the first phase and makes them available as a part of some production workflow. In this section, we'll discuss the following:
- Expectations
- Presentation options
- Development and testing
- Data quality management
Expectations
The primary goal of data science applications is to build "actionable" insights, actionable being the keyword. Many use cases such as fraud detection need the insights to be generated and made available in a consumable fashion in near real time, if you expect any action-ability at all. The end users of the data product vary with the use case. They may be customers of an e-commerce site or a decision maker of...
