Yakketo/analisis_supermarket
Proyecto de análisis de datos para predecir las ventas de un supermercado. Utiliza Python (Pandas, Scikit-learn, etc.) para el procesamiento de datos y la construcción del modelo, y Power BI para un dashboard interactivo que muestra el rendimiento del modelo y explora los datos de ventas.
What's novel
Proyecto de análisis de datos para predecir las ventas de un supermercado. Utiliza Python (Pandas, Scikit-learn, etc.) para el procesamiento de datos y la construcción del modelo, y Power BI para un dashboard interactivo que muestra el rendimiento del modelo y explora los datos de ventas.
Code Analysis
6 files read · 4 roundsGenerates synthetic Spanish supermarket sales data, cleans it, and trains a RandomForestRegressor with basic time-series features to predict units sold, saving results to CSV and a Power BI dashboard.
Strengths
The data generator is surprisingly detailed with realistic Spanish market factors (regional holidays, climate effects, seasonal product demand, price sensitivity). The code is well-organized with clear separation between data cleaning, feature engineering, and model training.
Weaknesses
The 'prediction' is in-sample (predicting on training data after lag-induced NaN removal), not a true out-of-sample forecast. Tests are basic smoke tests that only verify the pipeline runs without errors, and test_graficos.py is a plotting script masquerading as a test. The ML approach is a textbook sklearn pipeline with no novel methodology.
Score Breakdown
Signal breakdown
Innovation
Craft
Traction
Scope
Evidence
Commits
9
Contributors
2
Files
25
Active weeks
1
Repository
Language
Python
Stars
1
Forks
0
License
NOASSERTION