How hectorpartido/climateai works
ClimateAI is a Streamlit dashboard and an offline climate-priority modeling pipeline. Users explore country priorities through six screens; the shared dashboard layer loads saved results and invokes prediction, while the separate analysis command regenerates indicators, country groups, fuzzy priorities, and reports. The README documents screen capabilities beyond the sampled excerpts. Edges below are limited to calls, imports, or workflows visible in the supplied source index/excerpts or explicitly documented in the README.
Main entry points: comun.py Shared dashboard · train.py Analysis orchestrator · preprocess.py Indicator preparation · predict.py Priority prediction
Components 20
Dashboard
- Streamlit entry
app/streamlit_app.py - Shared dashboard
app/comun.py - Country panorama
app/paginas/1_panorama.py - Country ranking
app/paginas/2_ranking.py - Country profile
app/paginas/3_ficha.py - Country groups
app/paginas/4_grupos.py - Priority simulator
app/paginas/5_simulador.py - Data and model
app/paginas/6_datos_modelo.py
Analysis pipeline
- Indicator preparation
src/data/preprocess.py - Expert validation
src/data/hoja_experto.py - Analysis orchestrator
src/models/train.py - Country clustering
src/models/clustering.py - Analysis reports
reports - Model configuration
src/config.py
Model inference
- Fuzzy calibration
src/models/fuzzy.py - Saved models
models - Priority prediction
src/models/predict.py
Climate data
- Source datasets
data/raw - Processed indicators
data/processed
Other components
- Dashboard user
Connections 24
- Dashboard user → Streamlit entry: opens dashboard
README.md - Streamlit entry → Shared dashboard: sets up
app/streamlit_app.py - Country panorama → Shared dashboard: uses shared data
app/paginas/1_panorama.py - Country ranking → Shared dashboard: uses shared data
app/paginas/2_ranking.py - Country profile → Shared dashboard: uses shared data
app/paginas/3_ficha.py - Country groups → Shared dashboard: uses shared data
app/paginas/4_grupos.py - Priority simulator → Shared dashboard: requests prediction
app/paginas/5_simulador.py - Data and model → Shared dashboard: uses shared data
app/paginas/6_datos_modelo.py - Shared dashboard → Priority prediction: calls prediction
app/comun.py - Shared dashboard → Saved models: loads saved models
app/comun.py - Country groups → Indicator preparation: prepares group data
app/paginas/4_grupos.py - Data and model → Analysis orchestrator: runs analysis
app/paginas/6_datos_modelo.py - Analysis orchestrator → Indicator preparation: generates indicators
src/models/train.py - Indicator preparation → Source datasets: reads source data
README.md - Indicator preparation → Processed indicators: writes indicators
src/data/preprocess.py - Analysis orchestrator → Country clustering: trains groups
src/models/train.py - Analysis orchestrator → Fuzzy calibration: calibrates priorities
src/models/train.py - Analysis orchestrator → Analysis reports: writes evaluations
src/models/train.py - Analysis orchestrator → Saved models: saves model files
src/models/train.py - Priority prediction → Fuzzy calibration: runs fuzzy inference
src/models/predict.py - Priority prediction → Saved models: loads model artifacts
src/models/predict.py - Expert validation → Processed indicators: uses country indicators
src/data/hoja_experto.py - Country clustering → Processed indicators: uses indicators
src/models/clustering.py - Analysis orchestrator → Model configuration: uses paths and settings
src/models/train.py