docs: rewrite Gamma prompt with slide-by-slide instructions based on actual deck content

Author: William-Hill

docs/gamma-update-prompt.md

@@ -1,160 +1,247 @@
 # Gamma.app Slide Deck Update Prompt
 
 Paste the prompt below into Gamma.app to update the **AI-Powered Student Success Analytics** slide deck.
+The current deck has 13 slides. This prompt tells you exactly what to change on each slide.
 
 ---
 
-## Prompt
+## Overall Framing
 
-Update the existing "AI-Powered Student Success Analytics" slide deck for the CodeBenders Datathon submission. The platform has been fully built and deployed. Replace any draft/planned framing with current, delivered-state language. Apply the following changes slide by slide, then ensure visual consistency throughout.
+- The platform is **live and deployed** for **Bishop State Community College (BSCC)**, a historically Black community college in Mobile, Alabama (~4,000 students/year). It is **not** a multi-institution prototype.
+- Remove University of Akron and KCTCS as co-equal institutions. They may appear only as "future institution onboarding" examples.
+- Replace all "will build / planned / customize for your context" language with "built / delivered / live."
+- Add Bishop State's colors (navy and gold) as accent colors where appropriate.
 
 ---
 
-### Overall Framing Changes
-
-- The platform is no longer described as a multi-institution prototype. It is a **live, deployed system built for Bishop State Community College (BSCC)**, a historically Black community college in Mobile, Alabama (~4,000 students/year). It is designed to extend to other PDP institutions.
-- Remove or de-emphasize KCTCS as a co-equal institution. References to University of Akron and KCTCS can appear as "future institution onboarding" examples only.
-- Replace any "we will build" or "planned" language with "we built" and "delivered."
+## Slide-by-Slide Instructions
 
 ---
 
-### Slide: Title / Cover
+### Slide 1 — Title / Cover
+
+**Current:** "AI-Powered Student Success Analytics · Transforming higher education data into actionable insights that improve student readiness, retention, and institutional outcomes · Team CodeBenders · William Hill · Farron Rucker · Audrey Webb"
 
+**Change to:**
 - Title: **AI-Powered Student Success Analytics**
 - Subtitle: **Bishop State Community College × CodeBenders**
-- Add: "Live at [your-vercel-url]"
+- Below subtitle: `Live at [your-vercel-url]`
 - Tagline: *Turning PDP data into proactive student interventions*
+- Keep team names
 
 ---
 
-### Slide: Problem Statement
+### Slide 2 — The Challenge (currently "Three Institutions, One Critical Gap")
+
+**Current:** Three-column layout showing University of Akron, Bishop State Community College, and KCTCS as separate institutions with separate pain points.
 
-Replace with:
+**Replace entirely with:**
 
-**The Challenge at Bishop State Community College**
+**Title:** The Challenge at Bishop State Community College
+
+Replace the three-column institution layout with a single focused problem statement for BSCC:
 
 - 59% Black/African American student population — early intervention matters most for this community
-- 68% part-time enrollment — students juggling work, family, and school need proactive outreach
-- Gateway course bottlenecks in math and English are the #1 predictor of non-retention — but aren't surfaced in existing tools
-- Advisors lack unified, predictive views of student risk before academic difficulties compound
+- 68% part-time enrollment — students balancing work, family, and school need proactive outreach
+- Gateway course bottlenecks in math and English are the #1 predictor of non-retention — but not surfaced in existing tools
+- Advisors lack unified, predictive views of student risk before difficulties compound
 - PDP reporting is manual and annual — no mid-cycle alerting capability
 
+Remove all references to Akron and KCTCS from this slide.
+
 ---
 
-### Slide: Solution Overview
+### Slide 3 — The Common Thread (currently "Customizable Dashboards / Readiness Assessment / Predictive Analytics")
 
-**What We Built**
+**Current:** Three generic columns describing what institutions "need."
 
-A full-stack AI analytics platform with three layers:
+**Replace with:**
 
-1. **ML Pipeline** — 7 predictive models trained on 4,000 Bishop State students (retention, at-risk, gateway math/English success, GPA risk, time-to-credential, credential type)
-2. **Readiness Engine** — PDP-aligned rule-based scoring (academic 40% + engagement 30% + ML risk 30%) with full traceability and human-readable explanations
-3. **Live Dashboard** — Natural-language query interface, KPI tiles, retention risk charts, prompt history with re-run and audit trail
+**Title:** What Bishop State Needed — What We Built
+
+Convert the three columns to reflect delivered capabilities:
+
+| What They Needed | What We Delivered |
+|------------------|-------------------|
+| Quarterly refreshed dashboards with exportable insights | Live dashboard deployed on Vercel — KPI tiles, charts, export to CSV/JSON/Markdown |
+| Student readiness levels before they impact retention | PDP-aligned Readiness Index (Academic 40% + Engagement 30% + ML Risk 30%) scoring all 4,000 students |
+| Predictive models based on historical data | 7 ML models trained on 4,000 Bishop State students (XGBoost + Random Forest + Logistic Regression) |
 
 ---
 
-### Slide: Architecture / Tech Stack
+### Slide 4 — The Cost of Inaction (22.3% / $10.7B / 39%)
 
-Update the architecture diagram to reflect:
+**Current:** Three national statistics about dropout rates.
 
-- **Data Layer:** Bishop State PDP cohort + AR files → Python ML pipeline → Postgres (Supabase, hosted, US East)
-- **ML Layer:** XGBoost + Random Forest + Logistic Regression, 7 models, scikit-learn
-- **Application Layer:** Next.js 16 + React 19 + TypeScript, deployed on Vercel
-- **AI Features:** OpenAI-powered NLQ → SQL → Recharts visualizations
-- **Audit:** Server-side JSONL query log, prompt history in localStorage
+**Keep the national statistics** (they are still valid). Add one callout below the stats:
 
-Stack badges: Python · XGBoost · scikit-learn · Next.js · Supabase · Vercel · OpenAI
+> *At Bishop State, 68% of students enroll part-time — the demographic most impacted by these national trends and the most responsive to early intervention.*
 
 ---
 
-### Slide: The 7 Predictive Models
+### Slide 5 — Our Solution (NLQ / Visualizations / Predictive Models / Export & Share)
+
+**Current:** Four numbered points describing the solution generically.
 
-| Model | Output | Algorithm |
-|-------|--------|-----------|
-| Retention Prediction | Probability + risk tier | Logistic Regression |
-| At-Risk Early Warning | URGENT / HIGH / MODERATE / LOW | Composite rule engine |
-| Gateway Math Success | Pass probability | XGBoost |
-| Gateway English Success | Pass probability | XGBoost |
-| First-Semester GPA Risk | Low GPA probability | XGBoost |
-| Time-to-Credential | Predicted years to completion | Random Forest Regressor |
-| Credential Type | Associate / Certificate / Bachelor | Random Forest Classifier |
+**Update each point:**
 
-Trained on 4,000 Bishop State students with cross-validation and overfitting checks.
+1. **Natural Language Queries** — Type questions in plain English. Our OpenAI-powered interface translates them into SQL and returns a chart + raw data table instantly. No technical expertise required.
+2. **Instant Visualizations** — Bar, line, and pie charts generated on-demand from the `student_level_with_predictions` view. All charts include a raw data table for verification.
+3. **Predictive Models** — 7 ML models trained on 4,000 Bishop State students covering retention, at-risk alerting, gateway math and English success, GPA risk, time-to-credential, and credential type.
+4. **Export & Share** — Download the full dashboard as CSV, JSON, or Markdown report with one click.
 
 ---
 
-### Slide: Readiness Score
+### Slide 6 — Predictive Analytics Models (currently "6 ML Models Working Together")
 
-**PDP-Aligned Readiness Index**
+**Current:** Title says "6 ML Models." Lists: Retention Prediction, At-Risk Warning System, Time to Credential, Credential Type Prediction, Course Success/GPA Prediction, Gateway Math & English Success Prediction (combined).
 
-Formula:
-> Readiness = (Academic × 40%) + (Engagement × 30%) + (ML Risk × 30%)
+**Change title to:** **7 ML Models Working Together to Predict Student Success**
 
-Tiers: 🟢 High ≥ 0.65 · 🟡 Medium 0.40–0.64 · 🔴 Low < 0.40
+**Replace the model list with:**
 
-Current Bishop State distribution:
-- High Readiness: 83.9% (3,355 students)
-- Medium Readiness: 16.1% (645 students)
+| # | Model | Output | Algorithm |
+|---|-------|--------|-----------|
+| 1 | Retention Prediction | Probability + risk tier | Logistic Regression |
+| 2 | At-Risk Early Warning | URGENT / HIGH / MODERATE / LOW | Composite rule engine |
+| 3 | Gateway Math Success | Pass probability | XGBoost |
+| 4 | Gateway English Success | Pass probability | XGBoost |
+| 5 | First-Semester GPA Risk | Low GPA probability | XGBoost |
+| 6 | Time-to-Credential | Predicted years to completion | Random Forest Regressor |
+| 7 | Credential Type | Associate / Certificate / Bachelor | Random Forest Classifier |
 
-Grounded in: PDP momentum metrics, CCRC Multiple Measures research, Bird et al. (2021) transparency in predictive analytics.
+**Update the Key Model Insights** bullet:
+- Remove "23 features" — replace with "31 features across demographics, academic prep, enrollment, course performance, and Year 1 outcomes"
+- Keep the academic placement levels insight
+- Add: "All 7 models trained with cross-validation and overfitting checks on 4,000 Bishop State students"
 
-Every score is fully traceable — no black box.
+**Remove the "Implementation Timeline"** section — it describes a future plan; the models are already running.
 
 ---
 
-### Slide: Dashboard Features
+### Slide 7 — Solution Architecture
+
+**Current:** Abstract flow diagram: "PDP Cohort, PDP Course, AR Data → ML Models (5 Predictive Models) → Processing Pipeline → Storage & API → Frontend Dashboard"
+
+**Update to reflect the actual stack:**
 
-**What Advisors & Leadership See**
+Flow (left to right): **Data → ML Pipeline → Database → API → Dashboard**
 
-- **KPI Tiles:** Overall retention rate, at-risk student count, average readiness score
-- **Charts:** Retention risk distribution, readiness breakdown, at-risk alert levels
-- **NLQ Query Interface:** Type a question in plain English → get a chart + data table
-- **Prompt History:** Every query logged with timestamp, re-runnable in one click
-- **Methodology Page:** Research citations, scoring formula, worked examples (Maria T. → 0.699 High; Jordan M. → 0.386 Low)
+- **Data:** Bishop State PDP cohort + AR files (4,000 students)
+- **ML Pipeline:** Python · XGBoost · Random Forest · Logistic Regression · scikit-learn — 7 models
+- **Database:** Postgres (Supabase, hosted, US East) — `student_level_with_predictions` view
+- **API:** Next.js 16 API routes (serverless, Vercel) — `/api/dashboard/kpis`, `/api/analyze`, `/api/execute-sql`
+- **Dashboard:** React 19 + TypeScript + Tailwind CSS + Recharts — deployed on Vercel
+
+Add stack badges below the diagram: `Python · XGBoost · scikit-learn · Next.js 16 · React 19 · Supabase · Vercel · OpenAI`
 
 ---
 
-### Slide: FERPA & Transparency
+### Slide 8 — How It Works (Connect / Ask / Get Insights)
 
-**Built for Institutional Trust**
+**Current:** Three steps described generically ("Connect Your Data / Ask Questions / Get Insights").
 
-- No PII transmitted to any LLM provider — only aggregate behavioral metrics (GPA group, completion rate, placement level)
-- Student GUIDs excluded from stored features
-- Every readiness score traceable to its inputs
-- Server-side audit log of all NLQ queries (JSONL)
-- Methodology page publicly accessible for advisor onboarding
+**Update step descriptions:**
 
-Complies with FERPA §99.31(a)(1) for legitimate educational interest use.
+1. **Connect Your Data** — Bishop State's PDP cohort and AR files feed a Python ML pipeline. 7 models score every student and upsert results to Supabase with zero duplicates on re-run.
+2. **Ask Questions** — Type natural language queries like *"Show retention rate by credential type"* or *"How many students are at urgent risk?"*. OpenAI translates them to SQL in real time.
+3. **Get Insights** — Receive instant bar, line, or pie charts plus a raw data table. Every query is logged to localStorage and a server-side JSONL audit trail for FERPA compliance.
 
 ---
 
-### Slide: Results & Impact
+### Slide 9 — Live Demo
+
+**Current:** Just a title slide ("Live Demo").
+
+**Replace with:**
+
+**Title:** Live Platform
+
+Add content:
+- URL: `[your-vercel-url]` (bold, large text)
+- Three callout cards:
+  1. **KPI Dashboard** — Retention rate, at-risk count, readiness score, completion rate — all live from Supabase
+  2. **NLQ Query Interface** — Type any question, get a chart + data table in seconds
+  3. **Methodology Page** — Full scoring formula, research citations, worked examples (Maria T. → High 0.699; Jordan M. → Low 0.386)
+
+---
+
+### Slide 10 — Measurable Return on Investment
+
+**Current:** Four rows: Early Intervention, Increased Retention, Resource Advocacy, Leadership Engagement.
 
-**Delivered for Bishop State**
+**Keep structure. Update text:**
 
+- **Early Intervention** — 4,000 Bishop State students are now scored daily across 7 risk dimensions. Advisors can identify URGENT-risk students before they drop.
+- **Increased Retention** — Data-driven interventions targeting the 83.9% High Readiness vs. 16.1% Medium Readiness populations enable earlier, more effective outreach.
+- **Resource Advocacy** — The Readiness Index is grounded in CCRC Multiple Measures research and Bird et al. (2021) — citations that strengthen grant applications.
+- **Leadership Engagement** — Exportable CSV/JSON/Markdown reports and a live dashboard give leadership presentation-ready data at any time.
+
+---
+
+### Slide 11 — Real-World Impact Scenarios (For Advisors / For Leadership)
+
+**Current:** Two columns with generic bullet points.
+
+**Update bullets:**
+
+**For Advisors:**
+- Readiness scores + at-risk alert levels (URGENT/HIGH/MODERATE/LOW) prioritize which students to contact first
+- Gateway math and English success predictions identify students who need support *before* they fail the course
+- Natural-language query interface — no SQL needed to pull custom cohort reports
+- Prompt history lets advisors re-run any past query in one click
+
+**For Leadership:**
+- KPI dashboard shows overall retention rate, at-risk count, average readiness score, and course completion rate at a glance
+- 7-model prediction suite covers every major student outcome metric required for PDP reporting
+- Methodology page documents every formula and research citation — ready for accreditation or grant review
+- Full audit trail of all queries for FERPA compliance review
+
+---
+
+### Slide 12 — Your Next Steps (currently "Customize / Engage / Pilot and Scale")
+
+**Current:** Describes onboarding steps for a *prospective* institution.
+
+**Replace entirely with:**
+
+**Title:** What's Live — What's Next
+
+**Delivered (checkmarks):**
 - ✅ 4,000 students scored across 7 prediction dimensions
-- ✅ Live dashboard deployed (Vercel + Supabase)
-- ✅ NLQ interface with prompt history and audit trail
+- ✅ Live dashboard deployed (Vercel + Supabase, US East)
+- ✅ NLQ interface with prompt history and server-side audit trail
 - ✅ PDP-aligned readiness engine with research citations
 - ✅ Methodology page with worked examples for advisor transparency
-- ✅ Deploy script for ongoing data refresh
+
+**Roadmap:**
+- 🔲 Role-based access (Advisor / Leadership / IR views)
+- 🔲 GitHub Actions for automated deploy on `main` push
+- 🔲 Multi-institution onboarding (University of Akron, KCTCS)
+- 🔲 Student roster table with per-student drill-down
+- 🔲 Dashboard filtering by cohort, term, and demographic attributes
 
 ---
 
-### Slide: Next Steps / Roadmap
+### Slide 13 — Closing ("Let's Transform Student Success Together")
+
+**Current:** Generic closing slide with tagline "The question isn't whether we can predict student success. The question is: what will we do with that knowledge?"
+
+**Keep the tagline.** Update the body:
+
+> We built a live, deployed AI analytics platform for Bishop State Community College — 4,000 students scored, 7 models running, advisors empowered. The platform is live today at `[your-vercel-url]`.
 
-- **CI/CD:** GitHub Actions for automated Vercel deploy on `main` push
-- **Multi-institution:** Onboard University of Akron and KCTCS using the same PDP-aligned pipeline
-- **Role-based access:** Advisor vs. Leadership vs. IR views
-- **Scheduled refresh:** Quarterly PDP pipeline re-runs
-- **Enhanced NLQ:** Demographic equity gap queries, cohort comparison
+Add callout cards for three key numbers:
+- **4,000** students scored
+- **7** predictive models
+- **83.9%** High Readiness
 
 ---
 
-### Design Notes for Gamma
+## Design Notes
 
 - Keep the existing color scheme and layout style
-- Use data callout cards for the key numbers (4,000 students, 7 models, 83.9% High Readiness)
-- The architecture slide should use a left-to-right flow diagram: Data → ML Pipeline → Supabase → Next.js/Vercel → User
-- The readiness score slide should visually show the three weighted components adding up to the final score
-- Add Bishop State's colors (navy and gold) as accent colors where appropriate
+- Add Bishop State navy and gold as accent colors on data callout cards
+- The architecture slide (Slide 7) should use a left-to-right flow diagram
+- The readiness score content (Slide 3) should visually show the three weighted components (40% + 30% + 30%) summing to the final score
+- Use data callout cards for the key numbers: 4,000 students, 7 models, 83.9% High Readiness