How dggdaniel/trabajo-de-grado-wilson-gaona works
This repository is a reproducible research workflow analyzing subjective well-being in Colombia’s ECV 2024 data. It consolidates and filters survey data, prepares leakage-controlled text, generates embeddings through either OpenAI or local Qwen, evaluates traditional statistical and machine-learning models, and produces thesis figures. The README documents the execution stages and major methods; the excerpts also show specific input/output behavior. Connections between stages follow that documented workflow, while internal relationships in unsampled modules are not inferred.
Main entry points: 02_generar_embeddings_small_api_openai.py OpenAI embeddings · 01_preparacion_datos_serializacion.py Serialized texts · 01_consolidacion_y_depuracion_ecv2024.py Consolidate and audit · 03_generar_embeddings_qwen3_local.py Local Qwen embeddings
Components 16
Data processing
- Consolidate and audit
01_procesamiento_datos/01_consolidacion_y_depuracion_ecv2024.py - Select variables
01_procesamiento_datos/02_seleccion_y_depuracion_variables.py
Text embeddings
- Serialize text
02_generacion_embeddings/01_preparacion_datos_serializacion.py - OpenAI embeddings
02_generacion_embeddings/02_generar_embeddings_small_api_openai.py - Local Qwen embeddings
02_generacion_embeddings/03_generar_embeddings_qwen3_local.py
Model evaluation
- Statistical reference models
03_modelos_estadisticos_y_ml/01_modelos_estadisticos_referencia.py - GroupKFold and holdout
03_modelos_estadisticos_y_ml/02_evaluacion_group_kfold.py
Thesis figures
- Figures 1–7
04_visualizacion_y_figuras/generar_figuras.py
Other components
- Researcher
- DANE microdata
- OpenAI API
- Processed datasets
01_procesamiento_datos/02_seleccion_y_depuracion_variables.py - Serialized texts
02_generacion_embeddings/01_preparacion_datos_serializacion.py - Embedding outputs
02_generacion_embeddings/03_generar_embeddings_qwen3_local.py - Evaluation results
03_modelos_estadisticos_y_ml/02_evaluacion_group_kfold.py - Thesis figures
04_visualizacion_y_figuras/generar_figuras.py
Connections 22
- Researcher → Consolidate and audit: runs pipeline
README.md - DANE microdata → Consolidate and audit: supplies survey data
README.md - Consolidate and audit → Select variables: precedes filtering
README.md - Select variables → Processed datasets: exports dataset
01_procesamiento_datos/02_seleccion_y_depuracion_variables.py - Processed datasets → Serialize text: feeds preparation
- Researcher → Serialize text: runs preparation
README.md - Serialize text → Serialized texts: writes serialized text
02_generacion_embeddings/01_preparacion_datos_serializacion.py - Serialized texts → OpenAI embeddings: provides text
02_generacion_embeddings/02_generar_embeddings_small_api_openai.py - Serialized texts → Local Qwen embeddings: provides text
README.md - Researcher → OpenAI embeddings: runs option
README.md - Researcher → Local Qwen embeddings: runs alternative
README.md - OpenAI embeddings → OpenAI API: requests embeddings
02_generacion_embeddings/02_generar_embeddings_small_api_openai.py - OpenAI embeddings → Embedding outputs: exports embeddings
02_generacion_embeddings/02_generar_embeddings_small_api_openai.py - Local Qwen embeddings → Embedding outputs: exports embeddings
02_generacion_embeddings/03_generar_embeddings_qwen3_local.py - Researcher → Statistical reference models: runs reference models
README.md - Researcher → GroupKFold and holdout: runs evaluation
README.md - Embedding outputs → GroupKFold and holdout: supports model comparisons
- Statistical reference models → Evaluation results: exports tables
README.md - GroupKFold and holdout → Evaluation results: exports evaluation
03_modelos_estadisticos_y_ml/02_evaluacion_group_kfold.py - Researcher → Figures 1–7: runs figure generation
README.md - Evaluation results → Figures 1–7: informs comparison figure
README.md - Figures 1–7 → Thesis figures: saves figures
04_visualizacion_y_figuras/generar_figuras.py