How 17prabhanshu/ctrl-shift-n-winpeat works
WIE is a workforce-intelligence platform that analyzes four unlinked labor-market and talent datasets in isolated lanes, then compares statistical findings rather than joining raw records. Its documented workflow loads, validates, versions, quality-checks, and cleans data; specialized market, skill, and talent analyses produce evidence for a governed registry and interactive Streamlit dashboard. The graph combines README-documented lane relationships with source-index and excerpt evidence for implementation links; some analytical internals and dashboard wiring were not sampled.
Main entry points: benchmark_engine.py Talent benchmarking · evidence_registry.py Evidence registry · loader.py Dataset loader · cleaner.py Data cleaning
Components 26
Data preparation
- Dataset loader
src/ingestion/loader.py - Schema validation
src/validation/schema.py - Dataset versioning
src/validation/data_version.py - Quality checks
src/validation/quality.py - Data cleaning
src/cleaning/cleaner.py - Salary parsing
src/cleaning/salary_parser.py - Experience parsing
src/cleaning/experience_parser.py - Feature engineering
src/feature_engineering/features.py
Analytical lanes
- Market analysis
src/market_intelligence/market_analysis.py - Skill NLP benchmark
src/nlp/nlp_benchmark.py - Skill intelligence
src/skill_intelligence/skill_engine.py - Skill graph
src/skill_intelligence/skill_graph.py - Skill signal index
src/skill_intelligence/skill_signal_index.py - Talent benchmarking
src/models/benchmark_engine.py - Probability calibration
src/calibration/calibrator.py - Model explanations
src/explainability/explainer.py - Robustness analysis
src/benchmarking/robustness.py - Model ensemble
src/ensembles/model_council.py
Evidence governance
- Evidence registry
src/agents/evidence_registry.py - Analytical agents
src/agents/agent_architecture.py - Knowledge graph
src/agents/knowledge_graph.py
Results interface
- Interactive dashboard
app/app.py
Other components
- Analytics jobs
- Data science jobs
- JDS traits
- SDS traits
Connections 23
- Analytics jobs → Market analysis: analyzed by
- Data science jobs → Skill NLP benchmark: analyzed by
- JDS traits → Talent benchmarking: benchmarked by
- SDS traits → Talent benchmarking: benchmarked by
- Dataset loader → Schema validation: validates
generate_and_run.py - Dataset loader → Dataset versioning: versions
generate_and_run.py - Dataset loader → Quality checks: checks
generate_and_run.py - Dataset loader → Data cleaning: cleans
generate_and_run.py - Data cleaning → Salary parsing: parses salaries
src/cleaning/cleaner.py - Data cleaning → Experience parsing: parses experience
src/cleaning/cleaner.py - Data cleaning → Feature engineering: engineers features
generate_and_run.py - Skill intelligence → Skill graph: builds
- Skill intelligence → Skill signal index: indexes signals
- Talent benchmarking → Probability calibration: calibrates
src/calibration/calibrator.py - Talent benchmarking → Model explanations: explains
src/explainability/explainer.py - Talent benchmarking → Robustness analysis: tests robustness
src/benchmarking/robustness.py - Model ensemble → Talent benchmarking: loads benchmark data
src/ensembles/model_council.py - Market analysis → Evidence registry: registers findings
- Skill NLP benchmark → Evidence registry: registers findings
- Talent benchmarking → Evidence registry: registers findings
- Analytical agents → Evidence registry: audits claims
src/agents/agent_architecture.py - Analytical agents → Evidence registry: synthesizes claims
src/agents/agent_architecture.py - Evidence registry → Interactive dashboard: supplies evidence