Problem
Community health survey data collected in REDCap was siloed, inconsistently structured, and inaccessible to program staff who needed it for reporting and program improvement.
Learning Analytics & Data Pipeline
Data Engine — REDCap, Coupler.io & Looker Dashboard
A full-lifecycle data pipeline that transforms raw community health survey responses into decision-ready dashboards — turning messy data into meaningful insights for ODU stakeholders.
Screenshot — Looker Dashboard
Replace with an anonymized screenshot of the Looker Studio dashboard (blur or remove any sensitive data fields)
Project Snapshot
The core data challenge, the pipeline design, and the
stakeholder outcomes this project achieved.
Design Goal
Build an automated ETL pipeline that extracts survey data, cleans and transforms it for analysis, and surfaces key metrics in a live, shareable Looker dashboard.
Outcomes
- Real-time dashboard replacing manual monthly reports
- Stakeholders can filter by region, cohort, and time period
- Data refresh automated via Coupler.io — zero manual export
- Anomalies surfaced proactively via conditional formatting
The ETL Pipeline
How data flows from collection through transformation to the
final stakeholder-facing dashboard.
Infographic — ETL Pipeline Flowchart
Replace with a flowchart showing: REDCap → Coupler.io API pull → Google Sheets staging → Looker Studio dashboard
Extract — REDCap
REDCap serves as the primary data collection layer. Surveys are instrumented with field validation, branching logic, and required-field enforcement to minimize dirty data at the source.
- Survey branching mirrors instructional design scenario logic
- Field types enforced at entry (date pickers, dropdowns, validated text)
- API access configured for automated downstream pulls
Transform — Coupler.io
Coupler.io acts as the middleware layer, scheduling automated pulls from the REDCap API and mapping fields into a clean, analysis-ready schema in Google Sheets.
- Scheduled refresh every 24 hours (configurable)
- Column normalization and null-handling rules applied
- Staging sheet isolates raw data from calculated fields
Load & Visualize — Looker Studio
Looker Studio connects directly to the transformed Google Sheet, rendering live charts, filters, and scorecards that stakeholders can access without any technical knowledge.
- Interactive filters by program, cohort, date range, and geography
- Conditional formatting flags data quality issues inline
- Shareable via link — no login required for read-only access
Screenshot — Dashboard Detail View
Replace with an anonymized close-up showing chart types, filter controls, and conditional formatting in the live dashboard
Pipeline Flow
REDCap Survey → API pull via Coupler.io → Clean schema in Google Sheets → Live Looker Dashboard → Stakeholder decisions
Before This Pipeline
Staff manually exported CSVs from REDCap, cleaned data in Excel, and produced static PDF reports — a process taking 3–5 hours per reporting cycle.
After This Pipeline
Dashboard updates automatically. Staff spend zero time on export/clean/report cycles. Program leads access insights in real time between scheduled reporting periods.
ID Connection: Learning Analytics in Practice
Kirkpatrick Level 3 & 4 Measurement
This pipeline operationalizes Levels 3 (behavior transfer) and 4 (results) of the Kirkpatrick model. Rather than relying on post-training surveys alone, it connects community health behavior data directly to program activity data — enabling longitudinal outcome tracking that most L&D programs never achieve.
Data-Informed Program Improvement
The dashboard was used by program staff to identify knowledge gaps and under-served sub-populations — directly informing which content modules needed revision and which audiences needed additional outreach. Data became a design tool, not just a reporting artifact.