Diagnostics

Most analytical projects involve diagnosis, which is the process of finding causal or logical relationships between variables. Analyzing data uses mathematical transformations to find and validate relationships between variables. We might need to know whether complaints are clustered in a certain region of a service area or find the most likely cause of those complaints by relating it to other operational data.

This description implies that visualizing data to show a distribution or compare two or more variables is strictly speaking not analyzing anything. Descriptive statistics in most performance reports reduces the data to fewer numbers, but strictly speaking, does not add any information to the dataset. The average of a set of numbers or a trend line is already within the numbers. The defining property of diagnosis or analysis is that the cleaned data is transformed to reveal new information. An analysis shows us something that is not apparent from the data itself by combining...