Learning analytics:
measure what learners learn

Learning analytics uses data about learners and their environments to understand and improve learning. The Society for Learning Analytics Research defines it as the measurement, collection, analysis and reporting of data about learners and their contexts.¹ The work includes collecting data from learning management systems, assessments and xAPI activity records;² cleaning and combining datasets; defining meaningful metrics; identifying learners who need support; building dashboards for teachers, managers and L&D teams; evaluating courses and programmes; measuring impact on job performance; and handling data ethically and lawfully.

The field sits where education and data meet. L&D teams are under pressure to prove value — LinkedIn's Workplace Learning Report 2025 found 49% of learning leaders say executives worry about skills gaps.³ Data skills pay well in India; AmbitionBox lists data analysts at a typical ₹6.5–7.2 lakh, up 16% in two years.⁴ Learner data is personal, and India's DPDP Rules bring core obligations into force from May 2027,⁵ so privacy is part of the job.

Typical data analyst pay in India (AmbitionBox)

₹6.5–7.2 lakh

Two-year rise in data analyst pay (AmbitionBox)

16%

L&D leaders reporting executive skills worries (LinkedIn, 2025)

49%

When India's DPDP core duties apply

May 2027

Table: How learning analytics candidates are assessed (editorial summary)

Table: From engagement to impact (editorial summary)

Table: How learning analytics developed (editorial summary)

• An analysis of a sample learning dataset
• A dashboard for teachers or L&D managers
• A metric definition separating engagement and learning
• An impact evaluation plan
• A privacy and ethics checklist
• A verified learning analytics certificate

Show impact: "Analysed course data; found quiz gaming masking low mastery; redesign raised task pass rates; built teacher dashboard used weekly." Real outcomes only.

"Is completion rate a good measure?"

It shows persistence, not learning; pair it with assessment and performance measures.

"How would you identify struggling learners?"

Combine assessment performance, activity patterns and missed milestones, then involve teachers or mentors.

"How do you prove a training programme worked?"

Define outcomes, measure before and after, compare with a group that didn't attend and consider other factors.

"What are the risks of predicting dropouts?"

Bias, self-fulfilling labels and privacy; use predictions to offer support, not to exclude.

"How do you design a dashboard for teachers?"

Few metrics, clear actions — who needs help and on what.

• "What data sources are available?" — scope.
• "Who uses the analytics?" — audience.
• "How is learner consent handled?" — ethics.
• "What decisions should the data support?" — purpose.

Table: Learning analytics roles and typical pay in India (AmbitionBox, October 2026; indicative)

Table: Common learning analytics tools (editorial guidance)

• Education or L&D data executive — reports.
• Learning analyst — analysis and dashboards.
• Senior analyst / researcher — evaluations.
• Learning analytics lead — strategy.
• Head of learning insights — leadership.

1. Week 1: Data — Explore a public or sample learning dataset.
2. Week 2: Metrics — Define engagement and learning metrics and build a dashboard.
3. Week 3: Prove it — Plan an impact evaluation; take the Learning Analytics Certification.

Related: Assessment Design, EdTech and Data Visualization. Guides: assessment design, edtech and learning and development.

• Data pulls — LMS and assessments.
• Dashboards — weekly updates.
• Alerts — learners needing help.
• Evaluations — programme impact.
• Reviews — with designers and teachers.
• Governance — privacy checks.

"What's an LRS?"

A learning record store that stores xAPI statements about learning activities.

"What's mastery?"

Demonstrated ability to perform a skill to a defined standard.

Table: A first-90-days plan (editorial guidance)

Map data sources and fix quality issues. Replace vanity metrics with learning measures and build a useful dashboard. By the third month, run a first impact evaluation.

AI now powers adaptive learning paths, predicts at-risk learners and summarises open feedback. It can personalise support at scale, but predictions can be biased or opaque, and learners may feel watched. Analysts must test models for fairness, explain how predictions are used, and keep the focus on helping learners rather than ranking them.

Agree the business outcome with stakeholders first — sales, errors, safety incidents. Measure before training. Compare trained and untrained groups or use staggered rollouts. Account for other changes. Report the size of the effect and uncertainty honestly.

Identify signals that predict struggle — missed deadlines, low early scores, inactivity. Test which signals actually predict outcomes. Alert teachers or mentors with suggested actions. Review accuracy and fairness regularly.

Collect only data needed for learning purposes. Explain to learners what's collected and why. Limit access by role. Anonymise for research. Follow consent rules, especially for children.

Lead with the question the data answers and the main finding. Show one or two clear charts rather than a dense dashboard. Separate engagement measures from learning and impact measures. State limitations honestly, such as small samples or missing data. End with a recommended action, like redesigning a module or offering extra support to a group.

"More data means better insights." Relevant, valid data does.

"Time spent shows effort and learning." It may show confusion.

"Predictions are neutral." They can reflect past bias.

LMS exports are often messy: duplicate learner records, test accounts, inconsistent course names and timestamps in different time zones. Start by understanding what each field means and how it's generated. Remove test and staff accounts, merge duplicates using reliable identifiers, standardise course and module names, and convert timestamps to one time zone. Check for impossible values, such as completion before enrolment. Document every cleaning step so analyses can be repeated and trusted.

Agree definitions with the L&D or teaching team — for example, what counts as completion or an active learner — before building any dashboard.

Share findings as questions to explore together rather than verdicts. Ask teachers what the patterns mean in their classrooms and designers which content might cause them. Agree one or two changes to test, then measure again. Analytics improves learning only when the people who teach and design act on it.

Start from a decision someone needs to make, such as which module to redesign or which learners to support, and pick the smallest set of measures that informs it. Avoid tracking everything just because the data exists.

• Reporting completion as success.
• Dashboards without actions.
• No impact design.
• Ignoring privacy.
• Unchecked predictive models.

The definition comes from the Society for Learning Analytics Research; tracking from xAPI.com; India pay from AmbitionBox (read October 2026); learning priorities from LinkedIn's Workplace Learning Report 2025; and privacy timelines from India's DPDP Rules notification. The coding course scene, metrics table and history are editorial. Science of Apps publishes the free test linked here. Last reviewed October 2026.

• Learning analytics — using learner data to improve learning.
• xAPI — standard for recording learning activities.
• LRS — learning record store.
• Engagement metric — measure of activity.
• Impact evaluation — testing effects on performance.