Clean up package

Author: SamEdwardes

model/01-train-and-deploy-model/requirements.txt

@@ -6,4 +6,8 @@ scikit-learn==1.2.2
 vetiver==0.2.1
 ipykernel==6.23.1
 pyodbc==4.0.39
-git+https://github.com/sol-eng/bike_predict_python.git@samedwardes/bug-fixes#subdirectory=package
+
+# Install the `bikeshare` package. To bump the version install from a
+# new comment hash. E.g. update this string:
+# @4964b11c6e0bb5f70bfafc28c24234ce8c6b11f5
+git+https://github.com/sol-eng/bike_predict_python.git@fe23beb4f081a163cd30cbeede06808d4bc4decc#subdirectory=package

package/bikeshare/data.py

@@ -4,10 +4,7 @@
 def clean_data(df: pd.DataFrame) -> pd.DataFrame:
     out = df.copy()
 
-    # Clean the date column
-    out['date'] = pd.to_datetime(out['date'])
-
-    # Specify categories for dow
+    # Specify all possible values for day of week.
     days_of_week = [
         'Monday',
         'Tuesday',
@@ -17,12 +14,14 @@ def clean_data(df: pd.DataFrame) -> pd.DataFrame:
         'Saturday',
         'Sunday'
     ]
+
+    # Turn down into a categorical.
     out['dow'] = pd.Categorical(
         out['dow'],
         categories=days_of_week
     )
 
-    # One hot encode dow column
+    # One hot encode dow column.
     out = pd.get_dummies(out, prefix=[""], prefix_sep="", columns = ['dow'])
 
     # Select final columns