Add AI Opportunity layers: Advantage, Growth, and composite

[sidneyhori]

Summary

The existing "Digital AI Exposure" layer captures how much AI will reshape a job — but that single lens skews negative. A software developer scoring 9/10 feels alarming, even though demand for developers may actually grow. We wanted additional lenses to explore the positive side of AI's impact on occupations:

• AI Advantage (0–10): How much can a worker amplify their productivity and competitive edge by adopting AI tools? Scored via LLM with anchors from "minimal — work is physical/manual" (0) to "maximum — AI transforms every core task" (10).
• AI Growth (0–10): How much will AI expand demand, create new sub-roles, or grow the market for this occupation? Anchors range from "shrinking — AI directly replaces demand" (0) to "explosive — entirely new demand driven by AI" (10).
• AI Opportunity (0–10): Composite average of Advantage and Growth.

These complement rather than replace the existing exposure layer. Together, the four AI layers let you explore questions like: which occupations are highly exposed but also have high opportunity?

Changes

• score.py: Added --metric CLI flag (exposure | advantage | growth) with dedicated LLM prompts and separate output files per metric. Default (exposure) behavior is unchanged.
• build_site_data.py: Merges scores_advantage.json and scores_growth.json alongside existing scores.json, computes composite opportunity score. Gracefully handles missing files.
• site/index.html: 3 new layer toggle buttons, reuses the existing green↔red color scale (flipped: green = high = good), new stats dashboards, tooltips show all AI scores when hovering in any AI layer.
• make_prompt.py: Adds Advantage, Growth, Opportunity columns to the markdown report when score data is available.

Fully additive — zero changes to the existing exposure pipeline, UI, or data format.

Test plan

• uv run python score.py (no args) still works exactly as before
• uv run python score.py --metric advantage --start 0 --end 10 scores 10 occupations
• uv run python score.py --metric growth --start 0 --end 10 scores 10 occupations
• uv run python build_site_data.py merges all score files into site/data.json
• All 7 layer buttons render correctly in site/index.html
• Green-positive color scale for new layers (vs red-positive for exposure)
• Tooltips show all AI scores when hovering in any AI layer

Sidney Hori Hawthorne
X: @sid_hori
linkedin: https://www.linkedin.com/in/sidney-hori/
sidney@aeonixtech.com

🤖 Generated with Claude Code