LMS Integration Dashboard That Connected Siloed Systems and Gave Deans Student Performance Data They Couldn't Access Before

A private university's College of Arts & Sciences was operating with four disconnected systems that each held a piece of the student performance picture: Canvas LMS for coursework and grades, a separate Student Information System for enrollment and academic records, a standalone attendance tracking application, and an advising notes database maintained in SharePoint. Faculty, advisors, and department chairs each had access to one or two of these systems, but no one — including the dean's office — could see a unified view of how students were actually performing. At-risk students were identified late if at all, retention interventions were reactive, and data-driven decision-making was aspirational rather than operational.

JSG was engaged to build an integration layer that bridged all four systems into a unified analytics dashboard, giving department leadership real-time visibility into student performance and enabling early identification of at-risk students.

Client: A private university, College of Arts & Sciences

Employee Size: 65 faculty members

Industry: Higher Education

Services: - Data Integration & Pipeline Design - Analytics Dashboard Development - AI Automation & Workflow Design - Systems Integration

The College of Arts & Sciences served 2,200 students across 18 departments, and the disconnect between its information systems had been a known problem for years. Multiple committees had recommended integration, but the technical complexity and the lack of internal development resources had kept the project permanently on the "future" list.

First, the LMS and SIS operated as entirely separate systems with no data exchange. Canvas held real-time assignment grades, discussion participation, and course activity data, but the SIS — where official grades, enrollment status, and academic standing were recorded — could not access any of it until the end of the semester when faculty submitted final grades. This meant that the most current indicator of student performance — mid-semester coursework — was invisible to anyone outside the individual course shell. Advisors could not see whether their advisees were struggling in specific courses, and department chairs had no visibility into course-level performance trends across their programs.

Second, the attendance tracking system was a standalone web application that faculty used inconsistently. Some departments required it, others did not, and there was no integration with either the LMS or the SIS. Attendance data — one of the strongest predictors of student attrition — existed in isolation. When a student stopped attending classes, the information sat in the attendance system until someone happened to notice, which could be weeks later. There was no automated alert and no connection to advising workflows.

Third, advising notes were stored in a SharePoint list that advisors maintained manually. The notes contained critical contextual information — a student's family situation, work schedule, health issues, or academic concerns they had shared in confidence — but this information was accessible only to the individual advisor. When a student changed advisors, when a faculty member wanted to understand why a student's performance had dropped, or when the dean's office wanted to understand patterns across the college, the advising data was effectively locked away.

Fourth, the dean's office had been requesting a student performance dashboard for three consecutive academic years. Each request had been met with the same response: the data existed in four different systems, none of which could communicate, and the IT department did not have the resources to build a custom integration. The dean was making retention and resource allocation decisions based on end-of-semester grade reports that arrived too late to inform intervention and lacked the granularity to identify which students, courses, or departments needed attention.

Fifth, retention was a growing institutional priority. The college's first-to-second-year retention rate had declined from 81% to 76% over three years, and the provost had set a target of returning to 80%. Without integrated data, the college could not identify which students were at risk early enough to intervene, could not measure which interventions were effective, and could not allocate advising resources to the departments or populations where they were most needed.

The college needed a solution that would: - Integrate LMS, SIS, attendance, and advising data into a single accessible layer - Provide department chairs and deans with real-time student performance dashboards - Identify at-risk students early enough for meaningful intervention - Connect early-warning indicators to advising workflows - Support data-driven retention strategy at the college and department level

JSG designed and deployed a data integration pipeline and analytics dashboard that unified all four source systems, layered in AI-powered risk scoring, and connected early-warning alerts to the college's advising workflows.

Key Components

Data Pipeline Orchestration N8N workflows were configured to extract, transform, and synchronize data from Canvas LMS, the SIS, the attendance tracking system, and the SharePoint advising notes database on a nightly cycle — with key indicators (grade changes, missed classes, dropped courses) triggering near-real-time updates. The pipeline normalized data formats across systems, resolved student identity matching discrepancies, and loaded the unified dataset into an Azure SQL data warehouse purpose-built for analytics.

Student Performance Dashboards Power BI dashboards were developed for three distinct audiences. Department chairs received course-level performance views showing grade distributions, assignment completion rates, and attendance patterns across all sections in their department. The dean's office received college-wide views with drill-down capability by department, program, student cohort, and demographic. Individual advisors received a caseload view showing their advisees' performance across all courses in a single screen — something none of them had ever had access to before.

AI-Powered Early Warning Risk Scoring Azure OpenAI was integrated to generate a composite risk score for every student each week, drawing on LMS activity (login frequency, assignment submission patterns, grade trends), attendance records, and historical data from prior cohorts. The model was trained to weight indicators that the college's institutional research office had identified as most predictive of attrition: missed assignments in the first four weeks, attendance drop-off, and declining LMS login frequency. Students crossing configurable risk thresholds were flagged automatically.

At-Risk Student Alert Triggers N8N automation connected the risk scoring engine to the advising workflow. When a student's risk score crossed the intervention threshold, an alert was sent to the student's assigned advisor with a summary of the contributing factors — specific courses where performance was declining, attendance gaps, and any relevant advising notes from prior interactions. The advisor could then initiate outreach with full context, rather than a generic check-in.

Intervention Recommendation Engine Azure OpenAI generated suggested intervention strategies based on the student's risk profile and the outcomes of similar interventions with comparable students in prior semesters. Recommendations ranged from tutoring referrals and study group placement to financial aid counseling and reduced course loads. Advisors could accept, modify, or dismiss recommendations and log the intervention taken, creating a feedback loop that improved the model over time.

System Connectors Zapier connectors handled the bidirectional data flow between the LMS, SIS, and attendance system where direct API integration was not available. This included syncing enrollment changes from the SIS to Canvas, pulling grade book data from Canvas into the data warehouse, and pushing attendance records from the standalone tracking application into the unified dataset.

## Quantifiable Impact

• Student performance data visible to deans and department chairs in real time for the first time in the college's history
• At-risk students identified an average of 3.5 weeks earlier than under the prior end-of-semester review process
• 340 students flagged for early intervention in the first semester, with 78% receiving advisor outreach within 48 hours of the alert
• First-to-second-year retention rate improved from 76% to 79.2% in the first year — closing two-thirds of the gap to the provost's 80% target
• Data silos between LMS, SIS, attendance, and advising systems fully eliminated
• Advisor caseload visibility reduced the average time spent per student check-in from 25 minutes to 14 minutes by eliminating the need to log into multiple systems
• Department chairs reported saving an estimated 8 hours per month previously spent manually compiling performance data from separate sources

• Data Warehouse: Azure SQL
• Cloud: Microsoft Azure
• Analytics & Dashboards: Power BI (student performance dashboards, retention risk views, department-level reporting)
• AI & Automation: Azure OpenAI (early warning risk scoring, intervention recommendations)
• Workflow Orchestration: N8N (data pipeline orchestration, at-risk student alert triggers, advisor notification workflows)
• Integration: Zapier (LMS/SIS/attendance system connectors, bidirectional data sync)
• Source Systems: Canvas LMS, institutional SIS, standalone attendance tracker, SharePoint (advising notes)
• Communication: SendGrid (advisor alert notifications), Microsoft Teams (department chair weekly digests)