Causal analysis is emphatically not just about data; in causal analysis we must incorporate some understanding of the process that produces the data, and then we get something that was not in the data to begin with. But Fisher was right about one point: once you remove causation from statistics, reduction of data is the only thing left.
