04 · Projects and application
Move from individual concepts toward a complete modeling workflow. Three guided projects use real public datasets and run entirely in your browser: predict a house price, predict which customers will leave, and predict who survived the Titanic. Each has a Kaggle-style brief and can run on your own Kaggle download.
Available today
Helpful foundations: Python with pandas and scikit-learn, and the Algorithms lessons on regression and classification. Nothing is locked. New to this? Read how to handle an ML problem end to end first.
Predict a number: what a neighbourhood's homes are worth, and how wrong to expect to be.
Guided project · 10 steps · California housing data (20,640 districts) · Runs in your browser, notebook included.
Open House Price projectPredict a yes/no: who is likely to leave, and who to call when calls cost money.
Guided project · 10 steps · Telco customers (7,043) · Choose the cut-off with costs, then check fairness.
Open Customer Churn projectThe classic first Kaggle competition: who survived, and what a model adds beyond one common-sense rule.
Guided project · 10 steps · 891 passengers · Kaggle-style brief and submission file.
Open Titanic projectConnect preparation, training and evaluation in one exercise.
Existing Linear Regression activity · Follow a fictional housing problem through guided decisions. This is not the full House Price capstone.
Open guided housing projectComing soon
The documentation-quality project needs a human-reviewed dataset first, so it is not available yet.
Work toward classifying documents for human review.
Coming soon · Good / Needs Improvement labels will use a human-review and dataset process. Data and the full project are not available here.
Coming soon