if the inputs to a computational process are incorrect, then the outputs from the process will be incorrect.

if the analyst makes programming errors, no solution can be computed.

if the Garbage function of a machine learning algorithm is not properly configured, then the outputs will only show known data correlations.

information technology, so they are the result of computing processes.

a process of abstraction, so any data are the result of human decisions and choices.

empirical measurements, so they are the result of neutral observations of the world.

for which we cannot find any interpretation.

that has no structure at all.

 where each instance in the data set may have its own internal structure, and this structure is not necessarily the same in every instance.

otherwise they will not be able to work on their problems full-time.

they need to understand the business domain's problem, why it is important, and how a data science solution to the problem might fit into an organization's processes.

any data analytics technique and algorithms must be uniquely adapted to the domain problem in order to achieve correct computational solutions.