add prj 4

Author: Ed Rubin

.gitignore

@@ -3,7 +3,7 @@
 lecture/00[8-9]*
 
 # Temporary holds: Projects
-projects/project-00[4-9]*
+projects/project-00[5-9]*
 
 # Exams
 exam/exam*

README.md

@@ -216,6 +216,10 @@ Past, present, and future projects.
 
 [Help](projects/project-003/help-003.md)
 
+[**004** Prediction finale](projects/project-004)
+
+*Due:* Wednesday 19 March 2025 by midnight (before 11:59 PM) Pacific
+
 ## Class project
 
 [Outline of the project](https://github.com/edrubin/EC524W25/tree/master/projects/class-project)

projects/project-004/README.md

@@ -0,0 +1,13 @@
+# Prediction finale
+
+*Due: 19 March 2025*
+
+Today we return to the place where it all started: predicting housing-sales prices using a rich dataset on housing attributes.
+
+**Your job** Using everything you learned this quarter, build a model to predict housing prices. And beat your first project's model. And maybe even build the best model in the class.
+
+**What you submit** Submit a notebook to Canvas that includes all your fancy code *and* how your model ultimately performed. Remember, to get the model performance, you'll need to submit your predictions to Kaggle.
+
+**Hints** It might be helpful to review (or follow) [my Kaggle notebook](https://www.kaggle.com/code/edwardarubin/intro-tidymodels-split-kaggle) that shows you how to integrate pre-split data (like Kaggle's `train.csv` and `test.csv`) into the `tidymodels` framework.
+
+**How you're graded** You'll be graded on the quality of your model, your code, and whether it looks like you actually tried to beat your past score. So make sure to include all the code that led to your final model. (And make sure your model is good.)