Only a few times, but when I needed to solve something like that, it was really important. A few examples: processing a giant, highly-interconnected graph that

Only a few times, but when I needed to solve something like that, it was really important. A few examples: processing a giant, highly-interconnected graph that was too big for memory, so figuring out how to split it up into parts small enough to handle; learning Bayesian statistics to make strong inferences from millions of weakly-indicating data points; making inferences about the nature of edges in a graph with tens of millions of nodes.