How migueltr7/prueba_machi works
Operaciones Aeronáuticas Chile is a Kedro-based analytics and machine-learning project with two parallel tracks: a national aviation baseline and a Puerto Montt delay-risk proof of concept. The graph follows declared pipeline inputs and outputs; artifact locations are represented by the repository’s data-stage directories. Several subsystems are documented in the README but their implementation wiring was not sampled, so their internal details are intentionally kept broad.
Main entry points: retrasos_ml/nodes.py Join flights and weather · 01_raw Raw aviation data · 02_intermediate Preprocessed flights · ml/nodes.py Airport clustering and flight classification
Components 21
Kedro runtime
- Pipeline registry
src/operaciones_aeronaves/pipeline_registry.py - Data catalog
conf/base/catalog.yml
National baseline
- Preprocessed flights
data/02_intermediate - Integrated operations
data/03_primary - Baseline EDA
src/operaciones_aeronaves/pipelines/data_processing/nodes.py - Airport clustering and flight classification
src/operaciones_aeronaves/pipelines/ml/nodes.py - Baseline outputs
data/08_reporting
Delay-risk PoC
- Filter SCTE flights
src/operaciones_aeronaves/pipelines/retrasos_ml/nodes.py - Reconstruct delay target
src/operaciones_aeronaves/pipelines/retrasos_ml/nodes.py - Prepare weather features
src/operaciones_aeronaves/pipelines/retrasos_ml/nodes.py - Join flights and weather
src/operaciones_aeronaves/pipelines/retrasos_ml/nodes.py - Train delay model
src/operaciones_aeronaves/pipelines/retrasos_ml/nodes.py - Assess weather effect
src/operaciones_aeronaves/pipelines/retrasos_ml/nodes.py - Flight and day segments
src/operaciones_aeronaves/pipelines/retrasos_ml/segmentacion.py - Delay EDA
src/operaciones_aeronaves/pipelines/retrasos_ml/eda.py - Risk outputs
data/08_reporting
Data inventory
- Scan and profile raw files
src/operaciones_aeronaves/pipelines/data_inventory/nodes.py - Inventory report
data/08_reporting/data_inventory_report.md
Other components
- Analyst
- Raw aviation data
data/01_raw - Weather data
Connections 28
- Analyst → Pipeline registry: starts run
- Pipeline registry → Data catalog: loads datasets
- Raw aviation data → Data catalog: provides files
- Data catalog → Preprocessed flights: stores intermediate
- Preprocessed flights → Integrated operations: stores primary
- Raw aviation data → Baseline EDA: supplies flights
- Baseline EDA → Preprocessed flights: writes cleaned flights
- Preprocessed flights → Airport clustering and flight classification: supplies joined flights
- Raw aviation data → Airport clustering and flight classification: supplies airport data
- Airport clustering and flight classification → Baseline outputs: writes analyses
- Integrated operations → Baseline EDA: feeds plots
- Integrated operations → Airport clustering and flight classification: feeds models
- Raw aviation data → Filter SCTE flights: supplies flight log
- Filter SCTE flights → Reconstruct delay target: filters flights
- Reconstruct delay target → Join flights and weather: provides delay labels
- Weather data → Prepare weather features: provides weather
- Prepare weather features → Join flights and weather: provides features
- Join flights and weather → Train delay model: trains on joined data
- Join flights and weather → Assess weather effect: supports impact analysis
- Train delay model → Assess weather effect: provides fitted model
- Join flights and weather → Flight and day segments: feeds segmentation
- Join flights and weather → Delay EDA: feeds exploration
- Prepare weather features → Delay EDA: provides weather features
- Delay EDA → Risk outputs: writes report
- Assess weather effect → Risk outputs: writes impact results
- Flight and day segments → Risk outputs: writes segment results
- Raw aviation data → Scan and profile raw files: scans files
- Scan and profile raw files → Inventory report: summarizes inventory