I don't think we even need data on this. One or two day prior is unthinkable, just per common sense. Although data points would be valuable and appreciated but we don't really need data to understand the one or two day is just not enough time for anyone or anything.

It's not like there are a bunch of dedicated reviewers, it's always other researchers who review. I don't think some of the most valued papers like MapReduce or BigTable would be reviewed in just a day or two. It's just not a reasonable amount of time in any condition in terms of reviewing research papers.