It's always all about the data, right? That is because having a thorough understanding of your data can actually make or break the outcome of a project.

Perhaps of all the out of the box features that IBM Watson delivers, the most important may just be Refine. Refine means just that—to hone, enhance, polish, improve or even perfect—the data you are planning to use in your IBM Watson project.

But why would you modify your data? Well, there are various reasons you may want to consider (or possibly need to) modify the data you are using. For example:

• You might want to enrich the data by adding calculations. Let's consider patient data that includes the physical characteristics of each patient, such has sex, height, weight, age, and so on. If you are interested in how various physical characterizes effect life expectancies, you may want to add a calculation that determines the patient's individual body mass index (BMI), which is calculated using height and weight. This ensures a consistent way of calculating BMI and also saves time by eliminating the need to perform the calculation within each analysis or prediction.
• Very often you may want to focus on a particular subset of your data. This can be accomplished by setting up a data filter (for that particular area of your business).
• More commonly, an exercise to make your data more readable is to rename columns or change a data type.
• Based upon your requirements or interest in the data, you may want to modify the default aggregations.
• Finally, you might want to create hierarchies and groups within your data.