Scientific Python and Computing Toolkit

Purpose

This note tracks the technical stack I am building for computational mathematics and physics. The goal is not to list tools for their own sake, but to connect each tool to a kind of academic output: simulation, numerical analysis, plotting, data handling, machine learning experiments, symbolic or matrix-oriented computation, and polished reports.

Tool map

• NumPy: arrays, vectorisation, linear algebra, and clean numerical experiments.
• SciPy: numerical integration, optimisation, sparse matrices, differential equations, and eigenvalue problems.
• Matplotlib: reproducible scientific figures with explicit labels, units, and saved plotting scripts.
• Pandas: tabular data handling for experiment logs, parameter sweeps, and result summaries.
• PyTorch: tensor computation and later machine-learning or differentiable-modelling experiments.
• MATLAB: matrix-oriented numerical work and comparison with Python workflows.
• LaTeX: reports, mathematical exposition, beamer slides, and reproducible academic writing.

Connection to projects

Current projects make this toolkit concrete. The one-dimensional Schrodinger equation project needs finite differences, eigenvalue solvers, and plotting. The Gray-Scott project needs arrays, stable timestepping, parameter sweeps, and performance-aware implementation. The Lagrangian mechanics report uses LaTeX to turn a presentation into a durable written artifact.

Standards

For public work, code should be runnable from a clean checkout, figures should be generated from scripts rather than screenshots, and reports should state assumptions, numerical checks, limitations, and future work.