Task 15: Update methodology page — remove implementation refs, add worked example

[William-Hill]

Overview

The methodology page (codebenders-dashboard/app/methodology/page.tsx) currently exposes internal implementation details that are not useful (and potentially confusing) to end users like advisors and administrators. We also want to make the scoring math tangible with a concrete example.

Changes Required

1. Remove internal implementation references

Remove the following from the page header badges and the "Data Source" section:

• Badges: Remove Script: generate_readiness_scores.py and Table: llm_recommendations badges — these are internal details. Keep Version: rules_v1.
• "Data Source" section: Remove the entire section. It exposes internal table names (student_level_with_predictions, llm_recommendations, readiness_generation_runs), the script path, and the re-run command — none of which are meaningful to end users.

2. Add a worked example section

Add a new section titled "Worked Example" (placed after the Scoring Formula section, before FERPA) with two side-by-side examples — one High Readiness and one Low Readiness — to show the full range of outcomes.

---

Example A — High Readiness: Maria T.

First-generation student, Part-time enrollment

Input	Value
GPA (Year 1)	3.2
Course completion rate	0.83
Passing rate	0.78
Gateway courses completed	Math only (1 of 2)
Credits earned (Year 1)	9
Enrollment intensity	Part-time
Courses enrolled	5
Math placement	College-level (C)
Retention probability (ML)	0.72
At-risk alert	MODERATE

Academic sub-score (avg of 5):

• GPA: 3.2 / 4.0 = 0.80
• Course completion: 0.83
• Passing rate: 0.78
• Gateway: 0.5 + 0.25 = 0.75 (1 gateway done)
• Credit momentum: 9 credits → 0.60
• Academic = (0.80 + 0.83 + 0.78 + 0.75 + 0.60) / 5 = 0.752

Engagement sub-score (avg of 3):

• Enrollment intensity: PT → 0.50
• Courses enrolled: 5 / 10 = 0.50
• Math placement: C → 1.00
• Engagement = (0.50 + 0.50 + 1.00) / 3 = 0.667

ML sub-score (avg of 2):

• Retention probability: 0.72
• At-risk alert: MODERATE → 0.60
• ML = (0.72 + 0.60) / 2 = 0.660

Final score:

(0.752 × 0.40) + (0.667 × 0.30) + (0.660 × 0.30) = 0.301 + 0.200 + 0.198 = 0.699 → High Readiness ✓

---

Example B — Low Readiness: Jordan M.

Remedial-track student, Part-time enrollment

Input	Value
GPA (Year 1)	1.8
Course completion rate	0.55
Passing rate	0.50
Gateway courses completed	Neither (0 of 2)
Credits earned (Year 1)	4
Enrollment intensity	Part-time
Courses enrolled	3
Math placement	Remedial (R)
Retention probability (ML)	0.38
At-risk alert	HIGH

Academic sub-score (avg of 5):

• GPA: 1.8 / 4.0 = 0.45
• Course completion: 0.55
• Passing rate: 0.50
• Gateway: 0.50 (0 gateways done)
• Credit momentum: 4 credits → 0.30 (< 6)
• Academic = (0.45 + 0.55 + 0.50 + 0.50 + 0.30) / 5 = 0.460

Engagement sub-score (avg of 3):

• Enrollment intensity: PT → 0.50
• Courses enrolled: 3 / 10 = 0.30
• Math placement: R → 0.20
• Engagement = (0.50 + 0.30 + 0.20) / 3 = 0.333

ML sub-score (avg of 2):

• Retention probability: 0.38
• At-risk alert: HIGH → 0.30
• ML = (0.38 + 0.30) / 2 = 0.340

Final score:

(0.460 × 0.40) + (0.333 × 0.30) + (0.340 × 0.30) = 0.184 + 0.100 + 0.102 = 0.386 → Low Readiness

---

Acceptance Criteria

• Header badges no longer reference generate_readiness_scores.py or llm_recommendations
• "Data Source" section is removed
• New "Worked Example" section is present between Scoring Formula and FERPA
• Both examples (High and Low Readiness) are shown with full input tables and step-by-step math
• Final scores and tiers (High/Low) are visually highlighted with color
• Page still builds without TypeScript errors

Branch

rebranding/task-15-methodology-worked-example → PR into rebranding/bishop-state

[William-Hill]

Completed — PR #60 merged into rebranding/bishop-state.