GSoC 2025 Proposal Discussion – Bayesian Classification with PyMC

[Muhammad Shahzaib]

Dear NumFOCUS (PyMC) Mentors,

I hope this email finds you well. My name is Muhammad Shahzaib, a recent mathematics graduate with a specialization in differential geometry, and I’m excited to apply for Google Summer of Code 2025 with NumFOCUS under the PyMC sub-project. I’m a beginner in machine learning, currently learning through Google’s ML Crash Course and FreeCodeCamp tutorials, and I’ve completed university coursework in numerical analysis and mathematical statistics, which I believe aligns well with PyMC’s Bayesian focus.

I’m drafting a proposal for a small (~90-hour) project titled "Bayesian Classification of Synthetic Data with PyMC." The idea is to use PyMC to build a Bayesian logistic regression model for classifying synthetic data, preprocess features with NumPy (leveraging my linear algebra skills), and visualize posteriors with Matplotlib. This builds on my math background and interest in probabilistic modeling, inspired by concepts like black holes. I’d love your feedback on this idea—particularly on its scope, feasibility for a beginner, and alignment with PyMC’s goals.

I’ve attached a brief draft outline below and welcome any suggestions before submitting by April 8, 2025. Could you also advise on the best dataset or tools to focus on? I’m available via email (shahz...@gmail.com) or the NumFOCUS Google Group for further discussion. Thank you for your time and guidance!

Best regards,
Muhammad Shahzaib
LinkedIn: https://www.linkedin.com/in/muhammad-shahzaib-885837215

Draft Outline:

Goal: Develop a Bayesian classifier for synthetic data using PyMC.

Topics: Bayesian Statistical Modeling, Numerical Preprocessing, Result Visualization.

Methods: Generate synthetic data with NumPy, model with PyMC (MCMC sampling), visualize with Matplotlib.

Timeline: 10 weeks (~90 hours)—data prep (Weeks 1-4), modeling (Weeks 5-6), visualization/report (Weeks 7-10).

Deliverables: Code, report, visualizations.