Skip to content
View esiihle's full-sized avatar
4 followers · 4 following

Achievements

Achievement: YOLO

Achievements

Achievement: YOLO

Organizations

@Appointmate-software-design

Block or report esiihle

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
esiihle/README.md

Sicelwesihle Myeza

Computer Science graduate (University of the Witwatersrand) with first-class marks in mathematics, and a Data Analyst at iqbusiness. I turn data into calibrated probabilities and risk-aware decisions — and I'm building toward quantitative and credit-risk work.

What I'm building

  • v8-quant-football-model — a match-forecasting & staking engine: Dixon-Coles goal modelling with Bayesian shrinkage, probability calibration, fractional-Kelly staking, and walk-forward backtesting with closing-line value as the edge proof. In development.
  • credit-default-scorecard — a bank-style probability-of-default scorecard: WOE/IV feature engineering, logistic PD model, KS/Gini/AUC/PSI validation, PDO score scaling with reason codes, benchmarked against a gradient-boosted model. In development.
  • survey-overlap-qc & survey-compact-to-binary — Python tooling for survey-data quality control and reshaping compact responses into analysis-ready form.

Focus

Probabilistic forecasting · probability calibration · position sizing under uncertainty · credit-risk modelling (PD / scorecards) · backtesting & validation.

Toolkit

Python (pandas · NumPy · scikit-learn · statsmodels) · SQL · Power BI · Git · Jupyter

Foundation

BSc Computer Science, University of the Witwatersrand (2020–2024) — first-class marks in mathematics: probability, statistics, linear algebra.

Elsewhere

Portfolio · LinkedIn · sicelwesihlemyeza@gmail.com

Pinned Loading

  1. credit-default-scorecard credit-default-scorecard Public

    Python

  2. survey-compact-to-binary survey-compact-to-binary Public

    Codeframe-driven compact→binary survey conversion with built-in QC, nets, and round-tripping.

    Python

  3. survey-overlap-qc survey-overlap-qc Public

    Overlap/Venn variables and heavy-selector QC for binary survey data, with a self-contained HTML report.

    Python

  4. v8-quant-football-model v8-quant-football-model Public

    Python

  5. survey-wave-pipeline survey-wave-pipeline Public

    Wave-agnostic PySpark survey tracker pipeline: schema enforcement, window metrics, incremental runs (synthetic-data showcase).

    Python