fix: Possible fix to missing values

app.py

@@ -61,7 +61,7 @@ def plot_anomalies(df_test_value, data, anomalies):
 
 def clean_data(df):
     # Drop rows with any null data
-    df = df.dropna()
+    # df = df.dropna()
 
     # Check if the DataFrame already contains the correct columns
     if "timestamp" in df.columns and "value" in df.columns:
@@ -83,12 +83,12 @@ def clean_data(df):
         # Rename column
         df.rename(columns={"Hourly_Labor_Hours_Total": "value"}, inplace=True)
 
-    elif "Date_CY" in df.columns and "Hour" in df.columns and "Net_Sales_CY" in df.columns:
+   elif "Date_CY" in df.columns and "Hour" in df.columns and "Net_Sales_CY" in df.columns:
         # Convert "Date_CY" and "Hour" columns into datetime format
         df["timestamp"] = pd.to_datetime(df["Date_CY"]) + pd.to_timedelta(df["Hour"].astype(int), unit='h')
 
         # Handle the case where hour is 24
-        df.loc[df["timestamp"].dt.hour == 24, "timestamp"] = df["timestamp"] + pd.DateOffset(days=1)
+        df.loc[df["timestamp"].dt.hour == 24, "timestamp"] = df["timestamp"] - pd.DateOffset(days=1)
         df["timestamp"] = df["timestamp"].dt.floor('h')
 
         # Keep only necessary columns
@@ -97,6 +97,9 @@ def clean_data(df):
         # Rename column
         df.rename(columns={"Net_Sales_CY": "value"}, inplace=True)
 
+        # Drop rows where 'value' is NaN
+        df = df.dropna(subset=['value'])
+
         return df
 
     else: