Diogo Ribeiro

Data Science & Research Lead writing practical notes on statistics, machine learning, forecasting, and data engineering. The useful posts here should add analysis, examples, code, experiments, or field judgment rather than paraphrase search results.

Open Source Projects & Packages

This page collects software projects that are more useful as technical assets than as generic blog posts: installable packages, documentation, source repositories, examples and research tooling.

The Python projects below are published on PyPI under DiogoRibeiro7. They are grouped here by purpose so readers can see what each package is for before jumping into package documentation.

PyPI Packages

Package Current release Install Purpose
wifi-activity-recognition 0.2.0 pip install wifi-activity-recognition Human activity recognition using WiFi CSI and computer vision workflows.
tscv-vision 0.4.0 pip install tscv-vision NumPy-first structured representation engineering for time series.
pinnlab 0.6.1 pip install pinnlab Physics-informed neural network implementations.
setqca 0.2.0 pip install setqca Native Python toolkit for crisp-set and fuzzy-set qualitative comparative analysis.
gen-surv 2.0.0 pip install gen-surv Survival-data simulation and visualization for statistical research and benchmarking.
pinn-rk 0.6.0 pip install pinn-rk Runge-Kutta physics-informed neural networks with time-discrete losses in PyTorch.

Scientific Machine Learning

pinnlab

Physics-informed neural network implementations for experiments where differential equations, boundary conditions and neural approximators need to live in the same workflow.

pip install pinnlab

pinn-rk

Runge-Kutta PINNs for time-discrete physics-informed learning, including Gauss, Radau and Lobatto style losses in PyTorch.

pip install pinn-rk

Statistics, Survival Analysis and Research Methods

gen-surv

A Python package for simulating survival data and producing visualizations under Cox proportional hazards, accelerated failure time, multi-state, time-dependent covariate, hidden Markov, competing risks, mixture cure and piecewise exponential models.

pip install gen-surv

setqca

A native Python toolkit for crisp-set and fuzzy-set Qualitative Comparative Analysis. It belongs with the research-methods part of the site because it helps encode configurational arguments, not just fit predictive models.

pip install setqca

Time Series, Signals and Activity Recognition

tscv-vision

Structured representation engineering for time series with a NumPy-first API. This package is a better fit for reusable transformations and experiments than one-off notebook code.

pip install tscv-vision

wifi-activity-recognition

A package for human activity recognition using WiFi channel-state information and computer vision workflows. It sits at the intersection of sensing, signal processing and applied machine learning.

pip install wifi-activity-recognition

R Packages

myrpackage

A multilingual greeting and farewell package that serves as an example of proper R package structure.

Features:

  • Multilingual support (English, Spanish, French, Portuguese, German, Italian)
  • Comprehensive documentation with roxygen2
  • Full test coverage with testthat
  • Continuous integration with GitHub Actions
  • Proper package structure following R standards

View Documentation →

unconfoundedr

Test (un)confoundedness by comparing an effect from an RCT-like dataset to the same estimand from an observational dataset. Includes robust estimators, inference, and transportability tools.

Features:

  • IPW and AIPW (doubly robust) estimators for the marginal ATE
  • Bootstrap confidence intervals and Wald test
  • Transport modes: none, rct_to_obs, and auto (KS/energy shift detection)
  • Diagnostics for propensity overlap, stabilized weights, trimming, and transport ESS

View Documentation →

For source repositories, issues and development history, visit my GitHub profile.