Been working on a paper on still-emerging field and technology of learning analytics. Although last year’s Horizon Report gave it a “time-to-adoption horizon” of 2-3 years, it still has a way to go before there are products and findings that are reviewed and available in practice.
Both actual and potential applications of learning analytics can be represented through the diagram below, which maps out both the roles and functions that could be involved.
This paper, available here (pdf), also warns that the modest gains that could actually be made from learning analytics can easily be undermined by overselling it as a panacea for persistent educational challenges. It will not result in systems automatically customizing instruction to individual learning profiles and styles (if such things exist); but it could make blended and distance strategies a bit more response and reliable.
BTW: I would love to see this paper undergo peer review, and –as appropriate– publication for a wider audience!