How nahuelcedres/machine-learning-nahuel works
Project type and purpose
• This is a personal data-science project portfolio containing two Jupyter-notebook analyses:
a. “Customer Clustering – No supervisado” (unsupervised learning / clustering)
b. “My Personal Movie DataSet – Supervisado” (supervised learning / classification or regression)
• Purpose: Showcase data ingestion, preprocessing, modeling, evaluation, and reporting in notebook form.
File structure
• Top-level:
• “Customer Clustering – No supervisado” directory
• “My Personal Movie DataSet – Supervisado” directory
• README.md (portfolio overview)
• Each project directory has:
• Raw data files (.csv, .xlsx)
• Jupyter notebooks (.ipynb)
• Generated HTML reports (.html)
• Checkpoint folders (.ipynb_checkpoints)
README for additional insights
• README.md simply states that this is a personal portfolio of projects and ideas.
• No detailed architecture notes, so rely on the folder structure and standard data-science workflow.
a. Main components:
1. Data Storage
– Raw data files (.csv, .xlsx)
2. Processing & Analysis
– Jupyter notebook environment
– Python libraries (e.g., pandas, scikit-learn, matplotlib)
3. Model Artifacts
– In-memory models (within notebooks)
– No persisted serialized models in this structure
4. Reporting
– HTML exports of notebooks
5. Version Control / Portfolio Layer
– README.md
– Folder hierarchy managed in Git
b. Relationships / interactions:
• Data files → loaded by notebooks
• Notebooks → perform preprocessing → modeling → evaluation
• Notebooks → export results to HTML
• README.md → links to each HTML report for presentation
c. Architectural patterns / design principles:
• Linear data-science pipeline: ingestion → transform → model → report
• Separation of concerns by directory/project
• Reproducibility via notebooks and data files co-located
d. Relevant technologies:
• Jupyter Notebook
• Python ecosystem (pandas, numpy, scikit-learn, seaborn/matplotlib)
• Excel/CSV data sources
• Markdown/HTML for report rendering