Clean up week 04

Week 04/house_prices.Rmd

@@ -137,11 +137,7 @@ ggplot(data = data) +
 `SalePrice` and `LotArea`.
 
 ```{r}
-ggplot(data = data) +
-  geom_point(mapping = aes(x = LotArea, y = SalePrice), alpha = 0.3) +
-  geom_smooth(mapping = aes(x = LotArea, y = SalePrice), method = "loess") + 
-  scale_y_continuous(labels = comma) + 
-  scale_x_log10(labels = comma)
+# YOUR CODE GOES HERE
 
 ```
 
@@ -161,10 +157,7 @@ ggplot(data = data) +
 `SalePrice` and `OverallCond`.
 
 ```{r}
-ggplot(data = data) +
-  geom_boxplot(mapping = aes(y = SalePrice, x = OverallCond)) + 
-  geom_jitter(mapping = aes(y = SalePrice, x = OverallCond), color = "darkgreen", alpha = 0.1) + 
-  scale_y_log10(labels = comma)
+# YOUR CODE GOES HERE
 
 ```
 
@@ -185,8 +178,7 @@ stargazer(fit_01, intercept.bottom = FALSE, single.row = TRUE, type = "text")
 Create a simple linear regression model with `BedroomAbvGr` as the only explanatory variable.
 
 ```{r}
-fit_02 <- lm(SalePrice ~ BedroomAbvGr, data = data)
-stargazer(fit_02, intercept.bottom = FALSE, single.row = TRUE, type = "text")
+# YOUR CODE GOES HERE
 
 ```
 
@@ -195,8 +187,7 @@ stargazer(fit_02, intercept.bottom = FALSE, single.row = TRUE, type = "text")
 Create a multiple linear regression model with `GrLivArea` and `BedroomAbvGr` as explanatory variables.
 
 ```{r}
-fit_03 <- lm(SalePrice ~ GrLivArea + BedroomAbvGr, data = data)
-stargazer(fit_03, intercept.bottom = FALSE, single.row = TRUE, type = "text")
+# YOUR CODE GOES HERE
 
 ```
 
@@ -301,22 +292,6 @@ plotmo(fit_08)
 
 ```
 
-It's also possible to visually analyze interactions between explanatory variables, as we modeled them in model `fit_05`.
-
-The `level` argument controls which confidence (grey) and prediction (red) intervals should be plotted. The prediction interval, for example, tells us that, according to our model, 95% of the houses with a specific `GrLivArea` have a `SalePrice` within the shaded area. Read more about the difference between confidence and prediction intervals here: <http://www.sthda.com/english/articles/40-regression-analysis/166-predict-in-r-model-predictions-and-confidence-intervals/>.
-
-```{r}
-plotmo(fit_05, 
-       degree1 = c(),
-       #degree1 = c("GrLivArea"),
-       #degree2 = c(),
-       degree2 = c("GrLivArea", "LotArea"),
-       persp.ticktype="detailed",
-       level=.95,
-       theta=-35)
-
-```
-
 # Your Turn...
 
 Improve the above models, e.g., by including more variables and interaction effects. Try to find a balance between model fit and interpretability. Note: We are doing explanatory modeling here, not prediction! Hence, we are not working with a train/test split of the data.