Portfolio

A place on the web that gathers exercises, food, and habits you can rely on to deal with back pain. You can even subscribe to a kind mailing list to receive preselected suggestions with what you can focus on for the day/week/month.

An umbrella of websites as part of the Fairplay Sports organization. The goal? Providing team sport competitions among friends in a cheap, easy, and fun way! Using tools such as TypeScript, Django, Next.js, Tailwind, Render, and Vercel.

An implementation of constrained time-relaxed double round-robin (2RR) sports league scheduling using a tabu search based heuristic algorithm. A mouthful just to say a tool to do automatic scheduling of sports calendars.

A silly and simple random names or teams generator. Primarily an evening JavaScript practice, but also useful because wysiwyg and totally ad-free.

A scrollytelling web page that takes you on a journey through all released Arctic Monkeys albums. From song section importance to lyrical variation and lexical dispersion. Developed with the JavaScript library D3.js.

A handy application that acts as your digital futsal companion to help find a friendly opponent, look for a team to join, analyze vanity metrics, and receive tactical advice from a coachbot.

A Streamlit application that exposes an algorithm to detect the most similar cyclists given a selected rider, based on the race results over the past seasons. Partly built on a tiny AWS cloud infrastructure.

An R Shiny application to scan your job happiness. This is the password-protected development version. To see the wholesome beauty and utility of it, reach out to me so I can bring you in touch with Madame Fluks.

What breed is your pet? A simple tool that allows to detect the breed of your pet (actually, just cats or dogs), based on a single picture. Uses a computer vision AI model under the hood.

A more clunky predecessor of the cyclingsimilarity.com Streamlit application, with comparable functionality but using different data and Dash as framework. Deployed with Azure App Service.

The sentometrics R package (v1.0.0) offers a framework for textual sentiment time series analysis. It integrates the quantification of sentiment from texts, the aggregation into time series, and their usage for prediction.

Cobra is a Python library to build classification and regression models focused on performance and interpretation. I helped maintaining the package while at Python Predictions and being used in many actual client projects.

With this R Shiny application created during my PhD days you can compute both textual sentiment and aggregated time series. Enjoy fooling around with the various parameters!