Skip to content
View mwilczynska's full-sized avatar

Block or report mwilczynska

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
mwilczynska/README.md

Michael Wilczynska

Data scientist. Previously senior data scientist at Atlassian, and before that in NSW Government regulation and work health and safety research. PhD in astrophysics.

What is here

financial_datasets - daily asset-class series from 1970 to the present, built from primary sources. Each series documents its own method, records the chain joining modelled and proxy segments to later observed ones, and carries a source and a quality flag on every row. Fourteen per-asset methodology documents, a validation specification and per-dataset contract tests.

portfolio_optimisation - a brute-force grid search across the full weight space, backtested against that history. Reports CAGR, Sharpe, Sortino, Calmar, Ulcer index, maximum drawdown, VaR and CVaR, skew and excess kurtosis. The output is a set of portfolios from which the user can choose depending on their risk tolerance, time horizon and financial goals.

gmail-clean-and-backup - strips large attachments from a mailbox while keeping the messages, threading and labels intact. Every archived attachment is SHA-256 hashed on the way to disk, so it can be proved bit-identical months later. OAuth2, a seven-phase resumable commit, and a test suite under pytest, ruff and strict mypy.

Work related to my doctorate

The 253 Voigt profile models fitted for my thesis, and ecorrect, a command-line tool that scales continuum errors in quasar spectra. Two takes on robust outlier detection: a Fortran implementation of Rousseeuw's Least-Trimmed-Squares and an R comparison of methods. And cwhscolours, a small R package built at the Centre for Work Health and Safety.

Personal projects

travel-globe, a self-hosted travel blog with an interactive globe drawn from photo GPS · holiday-spend, a trip budgeting and spend tracker · cambio-score-tracker, a scorer for a card game.

Elsewhere

Google Scholar · LinkedIn

Sydney, Australia. Available from March 2027.

Pinned Loading

  1. financial_datasets financial_datasets Public

    Long-horizon daily asset-class datasets with reproducible Python pipelines, source provenance, and validation for portfolio backtesting.

    Python

  2. portfolio_optimisation portfolio_optimisation Public

    Brute-force portfolio grid search over daily asset histories back to 1970, with a dependency-free interactive results chart

    Python

  3. gmail-clean-and-backup gmail-clean-and-backup Public

    Reclaim Google storage by stripping large attachments from Gmail while keeping the emails, threading and labels intact. Attachments are archived locally with SHA-256 verification; originals stay re…

    Python

  4. holiday-spend holiday-spend Public

    A holiday spend tracking app.

    TypeScript

  5. Least-Trimmed-Squares-outlier-detection-and-regression Least-Trimmed-Squares-outlier-detection-and-regression Public

    Least-Trimmed Squares outlier detection and regression fitting

    Fortran

  6. PhD-Thesis-VPFIT-Voigt-profile-models PhD-Thesis-VPFIT-Voigt-profile-models Public

    253 VPFIT Voigt profile models from Michael Wilczynska's PhD Thesis