How marimar4/projet-python works
Project Type and Purpose
• This is a Python-based data-science workflow (not a web app or microservice).
• Goal: Empirically verify the Law of Okun by collecting, preparing, analyzing, and modeling socio-economic time-series data in Jupyter notebooks.
Key Components (to show
a. External Data Sources
• World Bank ZIP files:
– “GemDataEXTR.zip” (Unemployment, GDP deflator)
– “HNP_Stats_EXCEL.zip” (Demographic & health indicators)
b. Data Storage
• monmodule/bases (raw .zip archives, extracted Excel files, final_data.csv)
c. Data Ingestion & Preprocessing
• Notebook: Préparation_données.ipynb
• Python module: monmodule/declarations.py (utility functions, constants)
d. Exploratory Data Analysis (EDA)
• Notebook: EDA.ipynb
• Visualization libs: Pandas, Matplotlib, Seaborn, GeoPandas, Plotly (plotly.express)
e. Modeling
• Notebook: Modélisation.ipynb
• Statsmodels (time-series, transfer functions, ARDL)
f. Runtime Environment
• Jupyter Notebook server
• Python virtualenv defined by requirements.txt
g. Outputs
• Charts, statistical reports, model diagnostics
• final_data.csv
Interactions and Data Flow
• Data Storage → Data Ingestion & Preprocessing (reads raw files, cleans/merges data)
• Preprocessing → final_data.csv (intermediate artifact)
• final_data.csv & declarations → EDA (visualizations, descriptive stats)
• EDA → insights (charts)
• final_data.csv & declarations → Modeling (fit econometric models)
• Modeling → results (coefficients, validation plots)
Architectural Patterns and Design Principles
• Pipeline Pattern: sequential stages (Ingestion → EDA → Modeling)
• Modularization: utility functions in monmodule/declarations.py reused across notebooks
• Single Source of Truth: final_data.csv drives both EDA and Modeling
• Reproducibility: requirements.txt captures environment; notebooks document each step
Technologies and Frameworks
• Core: Python 3.x, Jupyter Notebook
• Data Handling: Pandas, NumPy
• Visualization: Matplotlib, Seaborn, Plotly, GeoPandas
• Modeling: Statsmodels
• Environment: requirements.txt, pipreqs