Apply suggestions from code review

Co-authored-by: Frits Hermans <post@fritshermans.nl>

Author: ArturoAmorQ

python_scripts/dimred_ex_01.py

@@ -12,7 +12,7 @@
 # ---
 
 # %% [markdown]
-# # Solution for Exercise M8.01
+# # Exercise M8.01
 #
 # In the Chapter on Linear Models we saw that feature engineering using
 # `PolynomialFeatures` can give a linear model the flexibility to capture

python_scripts/dimred_intuitions.py

@@ -125,7 +125,7 @@
 ax.bar_label(bars)
 ax.set_xlim([0, 14])
 ax.set_yticks([1, 2], labels=["PC1", "PC2"])
-ax.set_xlabel("eigenvalues")
+ax.set_xlabel("Explained variance")
 ax.set_ylabel("PCA features")
 ax.set_title("Variance Explained by Principal Components (PCA)", y=1.05)
 plt.show()

python_scripts/dimred_preprocessing.py

@@ -220,7 +220,7 @@ def plot_sq_loadings(ax, pca, feature_names, col_order=None, decimals=2):
 # with each other. Two strongly correlated features will jointly define a
 # direction with much higher variance than either one alone, and the explained
 # variance ratios across components will still be very unequal. Scaling removes
-# the unit bias; it does not make all directions equally important.
+# the magnitude bias; it does not make all directions equally important.
 #
 # There is a subtler issue too. `StandardScaler` estimates the standard
 # deviation from all samples, **including outliers**. `LotArea` (lot size in sq