← Back to portfolio

PYTHON / AUTOMATION / AI PROTOTYPES

Python automation, from notebooks to reusable workflows.

Independent, public-safe examples of data ingestion, CSV processing, source-grounded FAQ retrieval and technical documentation. No actual health records, private company datasets or production credentials are published.

Samsung Health CSV merger

A local Python and Pandas utility that reads Samsung export CSVs, handles a common extra-title-row format, retains source provenance and writes an error report. Mixed datasets remain distinct measurements; concatenation is not a time-aligned health analysis.

Explore the Python merger source →

Offline FAQ retrieval demonstration

An independently authored Python CLI searches fictional moon-library documents, returns verbatim answers with document and section attribution, and explicitly handles unsupported questions, ambiguous sources and retrieval limits. Automated tests run without AI services, credentials or model downloads.

Keyword ranking suggests relevant sources; only curated question wordings can produce an answer. This is a conservative retrieval example, with no generated responses or claim of semantic understanding.

Run the fictional FAQ demo and read its architecture →

Jupyter notebook review

Exercises progressed from variables and loops through functions, object-oriented programming, CSV processing, sentiment classification and FAQ chatbot prototypes. The published notes identify correctness issues and distinguish static review from model execution tests.

Read notebook review notes →

Reusable Python automation skill

A documented workflow for converting notebook experiments and CSV scripts into reproducible, privacy-conscious portfolio projects. Includes a sample CLI and a publication checklist. The GitHub skill is source documentation, not a hosted Python execution service.

View the reusable skill on GitHub →

Security note: All shared examples use generalised logic; source exports and personally identifying information remain private.