# Start by ignoring warnings for a smoother experience
import warnings
warnings.filterwarnings("ignore")

# Data setup: Load the monthly shampoo sales dataset
import pandas as pd
url = "https://raw.githubusercontent.com/jbrownlee/Datasets/master/shampoo.csv"
df = pd.read_csv(url)
df.columns = ["Month", "Sales"]
print("Data shape:", df.shape)
display(df.head())

# Plot the time series to see the sales over time
import matplotlib.pyplot as plt
plt.figure(figsize=(8, 4))
plt.plot(df["Month"], df["Sales"], marker="o")
plt.title("Monthly Shampoo Sales")
plt.xlabel("Month")
plt.ylabel("Sales")
plt.xticks(rotation=45)
plt.tight_layout()
plt.show()

# Creating a lag feature: sales from the previous month
df["Sales_Lag1"] = df["Sales"].shift(1)
display(df.head())

# Lag features for two and three months ago
df["Sales_Lag2"] = df["Sales"].shift(2)
df["Sales_Lag3"] = df["Sales"].shift(3)
display(df.head())

# Make a rolling average over 3 months
df["Rolling_Mean_3"] = df["Sales"].rolling(window=3).mean()
display(df.head(10))

# Compare sales, lag, and rolling mean on a plot
plt.figure(figsize=(10, 5))
plt.plot(df["Month"], df["Sales"], label="Actual Sales", marker="o")
plt.plot(df["Month"], df["Sales_Lag1"], label="Lag 1 Month", linestyle="--")
plt.plot(df["Month"], df["Rolling_Mean_3"], label="3-Month Avg", linewidth=3, alpha=0.7)
plt.xlabel("Month")
plt.ylabel("Sales")
plt.title("Sales, Lag, and Rolling Average")
plt.legend()
plt.xticks(rotation=45)
plt.tight_layout()
plt.show()

# Identify missing values caused by lagging and rolling
print(df.isnull().sum())

# Drop rows with missing values for modeling
df_clean = df.dropna().reset_index(drop=True)
print(df_clean.head())

# Example: Rolling sum and minimum sales for 3 months
df_clean["Rolling_Sum_3"] = df_clean["Sales"].rolling(window=3).sum()
df_clean["Rolling_Min_3"] = df_clean["Sales"].rolling(window=3).min()
display(df_clean.head())

# Iterating through rows: Print sales and lag for each month
for i, row in df_clean.iterrows():
    print(f"Month: {row['Month']}, Sales: {row['Sales']}, Lag-1: {row['Sales_Lag1']}")
    

# Quick feature creation with a list comprehension for lags 1-3
for l in range(1, 4):
    df_clean[f"Sales_Lag{l}_again"] = df_clean["Sales"].shift(l)
display(df_clean.head())

# Filtering: Find months where sales dropped below rolling mean
below_avg = df_clean[df_clean["Sales"] < df_clean["Rolling_Mean_3"]]
display(below_avg[["Month", "Sales", "Rolling_Mean_3"]].head())

# Mini project part 1: Predict the next month's sales using last month's sales (very simple!)
df_clean["Pred_Next_Month"] = df_clean["Sales_Lag1"]
df_clean["Actual_Next_Month"] = df_clean["Sales"].shift(-1)
df_proj = df_clean.dropna().reset_index(drop=True)
print(df_proj[["Sales", "Pred_Next_Month", "Actual_Next_Month"]].head())

# Mini project part 2: Compute mean absolute error (MAE)
mae = (df_proj["Pred_Next_Month"] - df_proj["Actual_Next_Month"]).abs().mean()
print(f"Mean Absolute Error: {mae:.2f}")

# Troubleshooting: What if you see lots of NaN values?
print(df["Sales_Lag1"].isnull().sum(), "missing values in Sales_Lag1")

# Extra tip: Try input() to create your own lag!
n = int(input("How many months back for lag? Enter a number: "))
df[f"Sales_Lag{n}_input"] = df["Sales"].shift(n)
display(df[["Month", "Sales", f"Sales_Lag{n}_input"]].head(n+3))