This chapter covers
- Assessing what you’ve learned in the preliminary phases of the project
- Revising the goals of the project based on new information
- Realizing when you should redo prior work
- Communicating new information to the customer and getting feedback
- Developing a plan for the execution phase
Figure 6.1 shows where we are in the data science process: beginning the build phase by formal planning. Throughout this book, I’ve stressed that uncertainty is one of the principle characteristics of data science work. If nothing were uncertain, a data scientist wouldn’t have to explore, hypothesize, assess, discover, or otherwise apply scientific methods to solve problems. Nor would a data scientist need to apply statistics—a field founded on uncertainty—in projects consisting only of absolute certainties. Because of this, every data science project comprises a series of open questions that are subsequently answered—partially or wholly—via a scientific process.