The French term mise en place is used in professional kitchens to describe the practice of chefs organizing and arranging the ingredients up to a point where it is ready to be used. It may be as simple as washing and picking herbs into inpidual leaves or chopping vegetables, or as complicated as caramelizing onions or slow cooking meats.
In the same way, before we start cooking the data or building a predictive model, we need to prepare the ingredients-the data. Our preparation covers three different tasks:
Loading the data into the analytic tool
Exploring the data to understand it and to find quality problems with it
Transforming the data to fix the quality problems
We say that the quality of data is high when it's appropriate for a specific use. In this chapter, we'll describe characteristics of data related to its quality.
As we've seen, our mise en place has three steps. After loading the data, we need to explore it and transform it. Exploring and transforming is an iterative process, but in this book, we'll pide it in two different steps for clarity.
In this chapter, we'll discuss the following topics:
Datasets and types of variables
Data quality
Loading data into Rattle
Assigning roles to the variables
Transforming variables to solve data quality problems and to improve data format of our predictive model
In this chapter, we'll cover how we explore the data to understand it and find quality problems.