Lucas Fonseca
  • Research
  • Projects & students
  • Resources
  • About & contact
Skip to content

DATA SCIENCE AND MACHINE LEARNING

Make the ideas move.

Explore the concepts behind Data Science and Machine Learning through small experiments. These activities grew out of materials I developed for my teaching and are open to everyone, whether or not you take the module.

Predict. Change one thing. Explain what happened.

Start with a question, make a prediction, then use the controls to test it. No coding, account or installation is needed. Each activity includes a numerical view and questions to help you make sense of the graphics.

Choose an experiment

01 / PREDICT A NUMBER

Fit a line or a curve

Watch gradient descent adjust a model. Compare linear and quadratic fits, inspect residuals and change the learning rate.

Explore: regression · loss · optimisation

Open regression activity

02 / DISCOVER GROUPS

Step inside k-means

Assign points, move centres and see clusters emerge in two or three dimensions. Try different starting points or work through a small example.

Explore: distance · centroids · initialisation

Open clustering activity

03 / PREDICT A CATEGORY

Compare decision boundaries

Give different classifiers the same points. Examine the boundaries they produce and compare training with held-out test performance.

Explore: classification · model complexity · generalisation

Open classification activity

New to machine learning?

Regression predicts a number. Classification predicts a category from examples with known labels. Both are forms of supervised learning. Clustering looks for structure in observations without using known category labels.

The activities use synthetic data so you can concentrate on the ideas. Start with regression, continue to clustering and finish with classification—or choose the question that interests you.

From an experiment to an explanation

For each activity, write down what you changed, what you expected and what you observed. Keep the data fixed when comparing settings. If the outcome surprises you, inspect the numbers as well as the plot.

The examples connect to questions about human movement and sensing, but their synthetic patterns are not experimental findings. Good performance in a small demonstration is a starting point for investigation.

Back to Resources · Explore student projects

These resources support exploration and revision. Any module-specific assessment instructions remain in the module’s usual teaching materials.

Lucas Fonseca
Human movement for rehabilitation, health and independence.

 

CHART · Email
University of Nottingham