First probabilistic framework for sequential random projection, an approach rooted in the challenges of sequential decision-making under uncertainty; A non-trivial martingale extension of Johnson-Lindenstrauss (JL) to sequentially adaptive data processes.