# Data setup
import warnings; warnings.filterwarnings("ignore")
import pandas as pd

url = "https://raw.githubusercontent.com/jbrownlee/Datasets/master/airline-passengers.csv"
data = pd.read_csv(url, parse_dates=['Month'])

print("Data shape:", data.shape)
print(data.head())

# Plot the time series to see its shape
import matplotlib.pyplot as plt

plt.figure(figsize=(10,4))
plt.plot(data['Month'], data['Passengers'], label="Passengers")
plt.xlabel('Month')
plt.ylabel('Passengers')
plt.title('Monthly Airline Passengers')
plt.legend()
plt.show()

# Calculate and plot Autocorrelation (ACF)
from statsmodels.graphics.tsaplots import plot_acf

plt.figure(figsize=(10,4))
plot_acf(data['Passengers'], lags=24)
plt.title('Autocorrelation (ACF) Plot')
plt.show()

# Plot Partial Autocorrelation (PACF)
from statsmodels.graphics.tsaplots import plot_pacf

plt.figure(figsize=(10,4))
plot_pacf(data['Passengers'], lags=24, method='ywm')
plt.title('Partial Autocorrelation (PACF) Plot')
plt.show()

# Print basic statistics
mean_value = data['Passengers'].mean()
max_value = data['Passengers'].max()
min_value = data['Passengers'].min()

print("Average passengers per month:", mean_value)
print("Maximum passengers in any month:", max_value)
print("Minimum passengers in any month:", min_value)

# Try accessing a column that does not exist to see a common error
try:
    print(data['Flights'])
except Exception as e:
    print("Error:", e)
    

# Add a calculated lagged column
data['Passengers_last_month'] = data['Passengers'].shift(1)
print(data[['Month','Passengers','Passengers_last_month']].head(10))

# Remove missing data created by shifting
data_lagged = data.dropna()
print("Old size:", data.shape, "New size:", data_lagged.shape)

# Compute the correlation between two months
cor = data_lagged['Passengers'].corr(data_lagged['Passengers_last_month'])
print("Correlation with previous month:", cor)

# Find the most highly autocorrelated lag up to 12 months
from statsmodels.tsa.stattools import acf
acf_values = acf(data['Passengers'], nlags=12)
print("Lag\tACF value")
for lag, val in enumerate(acf_values):
    print(lag, "\t", round(val, 3))
    

# Filter dataset to just the summer months as an example
data['Month_num'] = data['Month'].dt.month
summer = data[data['Month_num'].isin([6,7,8])]
print(summer.head())

# Plot the summer data for comparison
plt.figure(figsize=(8,4))
plt.plot(summer['Month'], summer['Passengers'], marker='o', linestyle='-')
plt.xlabel('Month')
plt.ylabel('Passengers')
plt.title('Summer Airline Passengers Only')
plt.show()

# Mini-project: Predict the next month's passengers using a simple average of last 12 months
def predict_next_month(df):
    last_12 = df['Passengers'][-12:]
    return int(last_12.mean())

prediction = predict_next_month(data)
print("Predicted passengers for next month:", prediction)

# Mini-project: Add a moving average column for trend smoothing
data['MA12'] = data['Passengers'].rolling(window=12).mean()

plt.figure(figsize=(10,4))
plt.plot(data['Month'], data['Passengers'], label="Actual")
plt.plot(data['Month'], data['MA12'], label="12-month MA", color='red')
plt.xlabel('Month')
plt.ylabel('Passengers')
plt.title('Passengers with 12-Month Moving Average')
plt.legend()
plt.show()

# Best practice: Always check for stationarity before predicting
from statsmodels.tsa.stattools import adfuller
result = adfuller(data['Passengers'].dropna())
print('ADF Statistic:', result[0])
print('p-value:', result[1])
if result[1] < 0.05:
    print("The series is likely stationary.")
else:
    print("The series is likely not stationary.")
    

# Troubleshooting: Handling missing values in time series
nan_count = data['Passengers'].isna().sum()
print("Number of missing values:", nan_count)
if nan_count > 0:
    print("Fill with previous value.")
    data['Passengers'] = data['Passengers'].fillna(method='ffill')
else:
    print("No missing values.")
    

# Extra tip: Explore multiple lags visually in a table
pd.set_option('display.max_columns', None)
lags = pd.concat([data['Passengers'].shift(i) for i in range(0, 4)], axis=1)
lags.columns = ['current', 'lag_1', 'lag_2', 'lag_3']
print(lags.head(10))

# Challenge: Ask the user for a lag and print its autocorrelation
user_lag = int(input("Enter a lag (number of months back, 1-12): "))
if 1 <= user_lag <= 12:
    correlation = data['Passengers'].autocorr(lag=user_lag)
    print(f"Autocorrelation at lag {user_lag}: {round(correlation, 3)}")
else:
    print("Please enter a lag between 1 and 12.")