While 'big-data' and 'large-datasets' are different things/concepts as mentioned in Midavalo's answer, I think such differences are not reflected in the current way questions are tagged big-data and large-datasets in GIS SE.
My opinion is that big-data is just being used for expressing 'large-datasets' and doing stuff in software with it. The theoretical part (analytical/science part) is completely missing from those questions.
This is not a surprise given our huge bias toward software/applied questions (for example, we even have a tag when the question is abstract --> Is it useful to have a [gis-principle] tag?).
Moreover, how are we going to draw the line between using one over another and making questions being tagged consistently, so tags become useful for searching both subjects?
Proposal:
Therefore, I vote for synonymizing tags (keep them unmerged for a few months so we can see how it goes) making large-datasets the master one. A new excerpt could be written to welcome both types of concepts.