Orange Data Mining for Teens: Visual Machine Learning From Zero to Capstone

No. Orange is visual programming software: you build analysis by placing widgets on a canvas and connecting them, so no code is required at any point in the 24 classes. Students who already know Python can go further using the Python Script widget, which runs real Python inside the workflow, but it is entirely optional.

Orange is open source machine learning and data visualisation software developed at the Bioinformatics Laboratory, Faculty of Computer and Information Science, University of Ljubljana. It is released under the GNU General Public License version 3 or later and is free to download and use on Windows, macOS and Linux. There is no subscription, no trial period, and no student licence to renew.

It is real machine learning. Orange runs standard algorithms including decision trees, k nearest neighbours, naive Bayes, logistic regression, random forests, gradient boosting, support vector machines and neural networks, and evaluates them with proper cross validation, ROC analysis and confusion matrices. The visual canvas changes how you assemble the pipeline, not what the pipeline does.

It is designed for teens aged 13 to 18. Orange itself requires no account and has no age restriction, so the age range reflects the reading level, the statistical reasoning and the project expectations of the curriculum rather than any restriction in the software. Younger students who are strong in maths can join after a free demo class where we check the fit.

Any reasonably modern Windows, macOS or Linux laptop or desktop that can install Orange, plus a stable internet connection for the live classes. Most classes run comfortably on modest hardware because the datasets are student sized. The image analytics class in Phase 4 also needs a working internet connection, since the image embedders run on a server.

The Python course teaches you to write the analysis in code with pandas and Python libraries, which means you spend the early weeks on syntax. This course removes the syntax barrier so students reach models, evaluation and clustering in the first month, and it goes deeper on honest evaluation and visual exploration. Many students take Orange first and then move to Python, arriving with the concepts already in place.

Yes, and Class 23 and Class 24 are built specifically for that. Students choose their own question, source a dataset responsibly, build a reproducible workflow, compare against a baseline, and export an Orange report as PDF for a teacher or judge, or as a .report file so another analyst can reopen the underlying workflows. The whole capstone is designed to be defended out loud.

Add-ons are optional widget packs installed from inside Orange through Options then Add-ons. This course installs and uses the Text add-on in Class 20 and the Image Analytics add-on in Class 21, and introduces the Educational, Explain and Fairness add-ons where they are relevant. They are free, like Orange itself, and we install them together in class.

Twenty four saved .ows workflows they can reopen and edit, a capstone project built from raw data to written conclusion, an exported report, a Modern Age Coders completion certificate, and, more usefully than any of those, the habit of asking how a model was tested before believing what it claims.