![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2431573%2F64c344ee5a32241a3dd4934fe48521eb%2Fkaggle_s.png?generation=1746113515183979&alt=media)

## Recommendations for the educators

- This competition can be used to teach data cleaning, EDA and regression techiniques. I used it for teaching model evaluation, metrics and hyperparameter tuning
- Go through the sample [notebook](https://github.com/STASYA00/AECinCode_tutorials/blob/main/notebooks/model_evaluation.ipynb) together with the students, making sure they understand how to download the data and upload the submission:)
- Give this task to the students already more or less comfortable with programming. Teaching programming on a kaggle competition proved to be a bad idea.
- Keep it short and sweet - max. 2 weeks

## Recommendations for the students

- Use the framework we have seen during the course to solve the homework (pandas, sklearn).
- Use Colab to write and run the code.
- [Notebook](https://github.com/STASYA00/AECinCode_tutorials/blob/main/notebooks/model_evaluation.ipynb) to get you started
- Save all the code you write on Github, then you will most certainly avoid the risk of losing your precious work.
- Organize your code following the approach shown during the lab sessions (e.g., dataset splitting, data loader, model creation, model fitting, etc) to ensure reproducibility of the solution.
- Use your domain knowledge to understand the data and the algorithms applied to them. Feel free to engineer more features that you feel can be relevant to the solution, even from the external sources.
- Feel free to write notebooks and tutorials in this competition explaining your approaches or writing utility code if you feel it might be useful for other participants. It will be considered during the course evaluation.
- Check what other people did with the [target dataset](https://www.kaggle.com/competitions/house-prices-advanced-regression-techniques/)

You don't have any constraint on the algorithm to be used, but try to explore different solutions. it will be appreciated during the final evaluation :)