Learning analytics is the discipline of collecting, measuring, and interpreting data about learners and learning contexts to understand what is happening and to improve outcomes. In a video-based platform the data sources include xAPI (Experience API) statements — actor-verb-object records such as "learner watched video to 80 %" — as well as quiz scores, navigation events, and time-on-task signals, all flowing into an LRS (Learning Record Store). From the LRS the data is queried by a reporting layer — a BI tool, a purpose-built dashboard, or a custom pipeline — that turns raw statements into actionable indicators for instructors and platform owners. The discipline borrows from educational data mining but is distinguished by its real-time or near-real-time orientation and its emphasis on intervention: the goal is not just to describe but to prompt a change, such as an instructor reaching out to a learner who has stalled. A key tension is between granularity and privacy: fine-grained video telemetry is powerful but also surfaces when, how often, and where a specific person struggled, which requires careful data governance and, under GDPR or FERPA, a documented legal basis. Learning analytics matures into value only when the metrics are tied to learning objectives rather than treated as raw activity logs. Without that alignment, high watch-time or completion numbers may obscure the fact that learners are not actually reaching mastery.
