Skip to content

curation_notes.md

Latest commit

 

History

History
62 lines (50 loc) · 2.75 KB

curation_notes.md

File metadata and controls

62 lines (50 loc) · 2.75 KB
Raw

Curation Notes

This repo was built by reviewing a much larger archive of personal Python files and selecting only the examples worth rewriting for public consumption.

Status Labels

Label Meaning
rewrite_keep Strong idea, rewritten for clarity and safety
coursework_keep Useful course example that stands on its own
archive_private Historically interesting but too rough, duplicative, or personal to publish
drop_throwaway Scratch work, duplicates, placeholder files, or low-signal experiments

Curation Rules

  1. No secrets, API keys, or tokens.
  2. No hardcoded local absolute paths.
  3. No publishable example should depend on personal brokerage automation.
  4. Duplicated model-iteration files get collapsed into one cleaner teaching version rather than ported as a family.
  5. If a script teaches an idea better as a clean rewrite than as a raw copy, rewrite it.
  6. If a script is too rough or low-signal, it stays in the classification inventory but does not enter the repo.

Files Rewritten Into This Repo

Source Destination
Legacy midpoint integration exercise calculus/riemann_sum_demo.py
Legacy PCA exercise linear_algebra/pca_walkthrough.py
Legacy Monte Carlo stock simulation finance/monte_carlo_gbm.py
Legacy correlation heatmap finance/correlation_heatmap.py
Legacy GARCH + Student-t workflow finance/garch_t_simulation.py
Big-data coursework Kafka/Spark examples streaming/*
LaTeX stripping helper text/latex_to_plaintext.py
SARIMA coursework final workflow time_series/sarima_workflow.py

What Was Deliberately Not Ported

  • Raw bot/trading automation scripts — not appropriate for a public educational repo without major legal and safety cleanup.
  • Brokerage-specific experiments — tied to specific APIs and credentials.
  • Redundant LSTM/hyperparameter branches — dozens of incremental model iterations with minimal conceptual separation; collapsed into nothing rather than picking one arbitrary snapshot.
  • Scripts with chaotic naming — files like stuf test 3.py, untitled1.py, etc. with no clear educational value.
  • Course submission artifacts — homework notebooks that are too tightly coupled to specific assignments to stand alone.

Classification Inventories

The complete file-by-file decisions are recorded in CSV format:

  • inventories/legacy_python_classification.csv — 150 files from ~/projects/legacy-python-scripts
  • inventories/standalone_python_classification.csv — 24 files from ~/Documents/Stevens-MS-Math and home-directory utilities
  • inventories/notebook_classification.csv — 12 Jupyter notebooks reviewed across all source locations