Learning Analytics

A deep dive into the four types of analytics used to understand and improve learning outcomes. 

I produced data visualizations, reference diagrams, and explainer videos that break down their purpose, methods, and tools.

Description of descriptive, diagnostic, predictive and prescriptive analytics

Descriptive Analytics

Visualizations and reference materials I created to illustrate how descriptive analytics summarizes past activity using basic measures, visual summaries, comparisons, and dashboards.

Basic Measures

Bar chart illustrating mean calculation with X-bar symbol across data points

Mean identifies average assessment scores across cohorts.

Bar chart illustrating median calculation highlighting the center value

Median provides a more accurate midpoint when outliers skewed results.

Bell curve showing standard deviation ranges from negative 3 to positive 3 sigma

Standard Deviation measures score spread to flag inconsistent learning outcomes.

Bell curve divided into percentile segments from 25th to 100th

Percentiles ranks learner performance to identify at-risk groups.

Tabular & Visual Summaries

Frequency table with columns for category, count, relative frequency, and cumulative frequency

Frequency Table organizes raw data into categories to reveal grade distributions at a glance.

Histogram visualizes score distributions to quickly spot patterns and gaps in learner performance.

Histogram showing score distributions by bin with frequency counts
Pivot table showing completion rate by cohort and module

Completion Rate by Cohort and Module uses pivot table analysis to pinpoint where specific cohorts dropped off.

Comparisons & Rankings

Line chart tracking month-over-month observed engagement trends

Month Over Month tracks engagement trends over time to identify seasonal drops or gains.

Line chart comparing year-over-year observed performance trends

Year Over Year compares annual performance to measure long-term program impact.

Pre/post comparison diagram showing score changes before and after training

Pre/Post Comparisons measure learning gains by comparing scores before and after training.

Horizontal bar chart ranking top learning items by completion count

Top N Ranking surfaces the highest-performing content to guide future course design.

Dashboards & Scorecards

Learning analytics dashboard with KPI cards, engagement timeline, and cohort comparison charts

Learning Analytics Dashboard consolidates key metrics into a single view for stakeholders to monitor program health.

Learning analytics dashboard with KPI cards, engagement timeline, and cohort comparison charts

Learning KPI Scorecard tracks cohort-level KPIs against targets to flag underperformance early.

Descriptive Analytics Video

Diagnostic Analytics

Diagrams and tool references illustrating how diagnostic analytics explains why outcomes occurred using drill-down analysis, correlation, and BI tools like Power BI and Tableau.

Techniques

Diagnostic analytics workflow from anomaly detection through drill down, correlation, regression, to root cause

Diagnostic Analytics Techniques illustrates a common diagnostic workflow moving from detecting anomalies to identifying root causes through drill down, pattern scanning, correlation, regression, and driver analysis.

Tools & Views

Power BI Key Influencers visual ranking factors that increase low completion likelihood

Key Influencers uses Power BI to surface which factors are most likely to increase the low completion rate, ranking each by strength of influence.

Root Cause Analysis combines completion trends with decomposition trees to trace performance drops back to specific cohorts, modules, and engagement patterns.

Predictive Analytics

Diagrams I created to map the predictive analytics workflow: how data moves through modeling methods and translates into scoring, forecasts, and risk flags.

Modeling & Forecasting

Predictive modeling workflow showing data flowing through statistical and machine learning methods to outcomes

Modeling and Forecasting illustrates how historical data flows through statistical and machine learning methods to generate predicted outcomes.

Predictive Outputs

Predictive outputs diagram showing three applications: scoring new data, forecasting trends, and flagging risk

Predictive Outputs shows three ways models deliver value: scoring new learners, forecasting future trends, and flagging at risk individuals for early intervention.

Prescriptive Analytics

This section will cover how prescriptive analytics recommends actions based on predictive insights and design judgment.

In Progress

Overview Video

I created this animated explainer to introduce the four analytics types and how they build on each other.