Multi-Year Teacher Performance Dashboards
(Tableau)

How interactive dashboards transformed multi-year assessment data into actionable insights for school leadership.

Snapshot Overview

  1. 👨‍💻My Role: Sole analyst

  2. 🔨Tools: Tableau, Excel

  3. 🔢 Data: Multi-year IA & GMAS assessment data (row-level, teacher-grain)

  4. 📊Records: 64,000+ student test scores | 400+ teachers | 4 school years

  5. 👥Stakeholders: District leadership, instructional coaching teams

  6. 🎯Outcome: Replaced manual spreadsheet review with interactive Tableau dashboards. Leadership conversations shifted from single-assessment scores to multi-year trends, so coaching support could be prioritized earlier and more precisely.

The Problem

Problem summary: Leadership had years of assessment data, but its format made trend analysis slow and subjective. The goal was an interactive tool for seeing teacher-level performance trends over time.

Georgia Cyber Academy, a K–12 online charter school, had multiple years of interim assessment (IA) and Georgia Milestones (GMAS) data across all grade levels. The data was complete, but it arrived as raw exports and large Excel pivot tables. Leadership could not quickly answer:

  • How individual teachers performed over time, not on a single assessment

  • Whether changes reflected meaningful trends or short-term fluctuation

  • How performance compared across grade bands, departments, and assessment types

  • Which teachers stood out, positively or negatively, without extensive manual filtering

Example of pre-aggregated reporting that masked longitudinal trends.

The Approach

Approach summary: Worked from row-level IA and GMAS data instead of pre-aggregated pivot tables, so performance could be tracked by teacher across school years.

  • Cleaned and standardized the data. Removed duplicate records and reconciled teacher names, course labels, and assessment names that changed across years and data sources.

  • Modeled it for comparison over time. Structured the data for year-over-year and assessment-to-assessment comparisons, with calculated fields to normalize scores across testing windows.

  • Built interactive Tableau dashboards filterable by grade band, department, course, and teacher.

  • Classified trends as increasing, stable, or declining so sustained changes stood out from short-term noise.

Raw assessment data required validation and standardization before longitudinal analysis was possible.

The Solution

I built two Tableau dashboards with the same trend logic: one for interim assessments (IA), given several times a year, and one for GMAS, given once a year. Keeping them separate made each easier to read.

Interim Assessment (IA) Dashboard

Shows how teacher performance shifts across assessment windows, within and across school years.

  • Assessment-to-assessment trends within each school year

  • Trend classification: increasing, stable, or declining

  • Filters for grade band, department, course, and teacher

Example of teacher-level IA trends highlighting short-cycle instructional response.

Design Iteration: From Density to Clarity

Early versions plotted hundreds of teacher trajectories at once and were hard to read. Later versions reduced visual noise and highlighted meaningful patterns without losing analytical depth.

Early iteration (high visual density)

Refined iteration (clarity & trend visibility)

GMAS Dashboard

Because GMAS is given once a year, this view focuses on year-over-year trajectories rather than short-term fluctuation.

Key features:

  • Multi-year trends at the teacher level

  • The same trend logic as the IA dashboard

  • Filters to compare grade bands and departments

Together, the two dashboards connect short-term instructional signals with long-term achievement outcomes.

Key Insights & Impact

Key Insights

  • IA trends did not always match GMAS outcomes. Teachers' short-term gains or declines across IA windows often looked different in their GMAS results, so IA trends are best read as context, not as predictors of summative results.

  • Trajectories told more than single scores. Teachers with similar scores on one assessment often had very different trajectories across assessment cycles.

  • Trends sharpened support decisions. Leadership could separate teachers with sustained declines from those with temporary dips, allowing more targeted coaching and resource allocation.

  • Context mattered. Patterns differed by grade band and department, which helped leadership avoid one-size-fits-all conclusions.

Impact

  • Leadership moved from manual spreadsheet review to visual, repeatable analysis

  • Conversations shifted from “How did this assessment go?” to “What is the trend over time?”

  • Data was used more consistently in instructional planning and coaching

  • The dashboards became a reusable framework for future assessment cycles

Reflection & What's Next

This project reinforced that analytics should be designed around decisions, not just reporting. The dashboards were built to prompt better questions and more targeted conversations, not to replace professional judgment.

Next steps would include:

  • Automated data ingestion with SQL-based transformations for near-real-time monitoring

  • Alerts for emerging negative trends

  • Additional context, such as enrollment changes or course placement, to explain variation in trends

  • More formal statistical methods to measure the strength and stability of trajectories

← All projects
Next project →