<project · intermediate>
House price predictor
Train and compare two regression models and explain which is better.
Build house price predictor one manageable milestone at a time. Keep your decisions, tests and finished work together so your portfolio shows what you learned, not only the final result.
- Regression
- Data leakage
- Evaluation
- Model comparison
Build it step by step
Choose data and target
Use a licensed public housing dataset and define the target currency and period. Identify features that would actually be available when predicting a new house.
Split before preprocessing
Reserve a held-out test set before imputing or scaling. Use a scikit-learn Pipeline and suitable random, geographic or time-based splits to avoid leakage.
Train two baselines
Start with a dummy mean predictor, then compare linear regression and a random forest. Keep training and cross-validation preprocessing inside each pipeline.
Evaluate fairly
Report MAE and RMSE on the same held-out examples. Plot residuals and inspect errors by price range without tuning the models against the test set.
Explain the choice
Save dependencies and reproducible training instructions. Discuss the trade-off between accuracy, interpretability and limitations; do not present this as professional valuation advice.
Share your finished work
Commit your source and a README explaining setup, decisions, tests and limitations. Remove secrets and personal data, credit dependencies and assets, then add the public repository or portfolio link and an optional working demo to LEARN Showcase. Hosting is optional; a clear repository or design case study is enough.
// done when
- Preprocessing is fitted only on training data
- Two regression models share the same evaluation split
- MAE and RMSE are reported
- A baseline and limitations are included
- README includes setup, credits and limitations; shared links are safe and accessible
// stretch goals
- Compare geographic splits
- Add permutation importance
Finished? Show it off
Add your project to the LEARN showcase so other learners and employers can see it.
Sign in to add your finished project to the showcase.