Advanced Machine Learning Projects in Python

This course is built for teens who can already write Python and want to understand machine learning properly, not just call a library and hope for the best. Over eight weeks, learners work through the full machine learning workflow: framing a problem, preparing data, choosing and training models, evaluating them honestly and communicating results.

The first half focuses on classic machine learning with pandas and scikit-learn. Learners explore and clean real datasets, build regression models to predict numbers, and train decision trees and random forests to classify data. They learn why models overfit, how cross-validation and train-test splits protect against it, and how to tune hyperparameters with care.

The second half moves into deep learning with Keras. Learners build a neural network from scratch, train a convolutional network to recognise images and try simple text classification. Throughout, the teacher explains the maths intuitively with diagrams, so learners understand what is happening inside each model without needing calculus.

In the final weeks, each learner chooses a dataset and builds a capstone project, then presents their approach, results and limitations to the group. They leave with a portfolio-ready notebook and the confidence to tackle new machine learning problems on their own.

How classes work

Each 75-minute class combines a short concept explanation, live coding in Google Colab and an extended hands-on challenge with teacher support. Optional practice notebooks are shared after each class for learners who want extra work.

What we'll cover

• 

  Week 1: The ML Workflow and Data with pandas – Frame a problem, load and explore a dataset and clean messy data.

• 

  Week 2: Regression and Evaluation – Predict numbers with linear regression and measure error properly.

• 

  Week 3: Classification with Trees and Forests – Train decision trees and random forests and read a confusion matrix.

• 

  Week 4: Overfitting and Model Tuning – Use train-test splits, cross-validation and hyperparameter tuning.

• 

  Week 5: Neural Networks with Keras – Build, train and visualise a simple neural network.

• 

  Week 6: Computer Vision with CNNs – Train a convolutional neural network to classify images.

• 

  Week 7: Working with Text – Turn text into features and build a simple text classifier.

• 

  Week 8: Capstone Project and Presentations – Finish an independent project and present results and limitations.

Learners should be confident with Python (functions, loops, lists and dictionaries) and comfortable with algebra. Prior exposure to pandas helps but is not required. A free Google account is needed for Colab.