Which of the following questions cannot be answered after analyzing data using data mining software?

• A. What is the trend?
• B. What is the correlation?
• C. How is our data set collected?
• D. What is the average sale?

Answer: (C)

Analyzing data with mining software reveals patterns and summaries that come directly from the values in the dataset. You can see trends by looking at how a metric changes over time, you can measure how two variables relate by computing a correlation, and you can summarize a single measure like the average sale to understand typical values. These are all derived from the data itself and the computations the software performs on it.

What you can’t determine from that analysis alone is how the data set was collected. Details about data provenance—where the data came from, the sampling method, the instruments or surveys used, the time frame, and any inclusion or exclusion criteria—aren’t embedded in the values you analyze. To answer how the data were collected, you need metadata or documentation that describes the data’s origin and collection process, not just the numbers produced by the analysis.