Hotel Booking Demand

import pandas as pd
bookings = pd.read_csv("hotel_bookings_clean.csv")
print(bookings.shape)
bookings.head(100)

list(bookings.columns)

bookings['Family_size']= bookings['adults']+ bookings['children']+bookings['babies']
cancellation_rate = bookings.groupby('Family_size')['is_canceled'].mean().sort_values(ascending= False)
print(cancellation_rate)

import matplotlib.pyplot as plt
import seaborn as sns
month_order = [
    'January', 'February', 'March', 'April', 'May', 'June',
    'July', 'August', 'September', 'October', 'November', 'December'
]
bookings['arrival_date_month']= pd.Categorical(bookings['arrival_date_month'], categories= month_order, ordered = True)


# Calculate cancellation rates by month
Monthly_cancellation = bookings.groupby('arrival_date_month')['is_canceled'].mean().reset_index()

Monthly_cancellation.head()

# First, verify your data
print(Monthly_cancellation.head())
print("\nDataFrame shape:", Monthly_cancellation.shape)
print("Null values in 'is_canceled':", Monthly_cancellation['is_canceled'].isna().sum())

# Robust version of the plotting code
plt.figure(figsize=(12, 6))

# Check if we have valid data to plot
if len(Monthly_cancellation) > 0 and not Monthly_cancellation['is_canceled'].isna().all():
    sns.lineplot(
        data=Monthly_cancellation,
        x='arrival_date_month',
        y='is_canceled',
        marker='o',
        color='royalblue',
        linewidth=2.5
    )
    
    plt.title('Monthly Hotel Booking Cancellation Rates', fontsize=16, pad=20)
    plt.xlabel('Month', fontsize=14)
    plt.ylabel('Cancellation Rate', fontsize=14)
    plt.xticks(rotation=45, ha='right')
    plt.grid(axis='y', linestyle='--', alpha=0.4)
    plt.gca().yaxis.set_major_formatter(plt.matplotlib.ticker.PercentFormatter(1.0))
    
    # Only add peak line if we have valid data
    if not Monthly_cancellation['is_canceled'].isna().any():
        peak_month = Monthly_cancellation.loc[Monthly_cancellation['is_canceled'].idxmax()]
        plt.axvline(
            x=peak_month['arrival_date_month'],
            color='crimson',
            linestyle='--',
            linewidth=1.5,
            label=f'Peak: {peak_month["arrival_date_month"]} ({peak_month["is_canceled"]:.1%})'
        )
        
        # Add value labels
        for index, row in Monthly_cancellation.iterrows():
            plt.text(
                x=row['arrival_date_month'],
                y=row['is_canceled'] + 0.01,
                s=f'{row["is_canceled"]:.1%}',
                ha='center',
                fontsize=10
            )
        
        plt.legend()
    
    plt.tight_layout()
    plt.show()
else:
    print("No valid cancellation data available to plot.")

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

# Ensure data quality
bookings = bookings.dropna(subset=['arrival_date_day_of_month', 'is_canceled'])

# Calculate cancellation rates by day
daily_cancel = (bookings.groupby('arrival_date_day_of_month')['is_canceled']
                .mean()
                .reset_index())

# Create the plot
plt.figure(figsize=(14, 6))

# Line plot with markers
ax = sns.lineplot(
    data=daily_cancel,
    x='arrival_date_day_of_month',
    y='is_canceled',
    marker='o',
    color='teal',
    linewidth=2.5
)

# Customize the plot
plt.title('Daily Cancellation Rates (Day of Month)', fontsize=16, pad=20)
plt.xlabel('Day of Month', fontsize=14)
plt.ylabel('Cancellation Rate', fontsize=14)
plt.xticks(range(1, 32))  # Show all days 1-31
plt.grid(axis='y', linestyle='--', alpha=0.4)

# Format as percentages
ax.yaxis.set_major_formatter(plt.matplotlib.ticker.PercentFormatter(1.0))

# Highlight peak day if data exists
if not daily_cancel.empty:
    peak_day = daily_cancel.loc[daily_cancel['is_canceled'].idxmax()]
    plt.axvline(
        x=peak_day['arrival_date_day_of_month'],
        color='coral',
        linestyle='--',
        label=f'Peak: Day {peak_day["arrival_date_day_of_month"]} ({peak_day["is_canceled"]:.1%})'
    )
    
    # Add data labels
    for _, row in daily_cancel.iterrows():
        ax.text(
            x=row['arrival_date_day_of_month'],
            y=row['is_canceled'] + 0.01,
            s=f'{row["is_canceled"]:.1%}',
            ha='center',
            fontsize=9
        )

plt.legend()
plt.tight_layout()
plt.show()

correlation = bookings['lead_time'].corr(bookings['is_canceled'])
print(f"Correlation between lead time and cancellations: {correlation:.2f}")

plt.figure(figsize=(10, 6))

# Convert is_canceled to strings for the palette mapping
sns.boxplot(
    data=bookings,
    x='is_canceled',
    y='lead_time',
    palette={'0': 'green', '1': 'red'}  # Keys as strings!
)

plt.title('Lead Time Distribution: Canceled vs. Not Canceled')
plt.xlabel('Canceled (1 = Yes, 0 = No)')
plt.ylabel('Lead Time (Days)')
plt.show()

# 2. Cancellation rate by lead time bins
bookings['lead_time_bin'] = pd.cut(
    bookings['lead_time'],
    bins=[0, 30, 60, 90, 180, 365, float('inf')],
    labels=['0-30', '30-60', '60-90', '90-180', '180-365', '365+']
)

bin_rates = bookings.groupby('lead_time_bin')['is_canceled'].mean().reset_index()

plt.figure(figsize=(12, 6))
sns.barplot(
    data=bin_rates,
    x='lead_time_bin',
    y='is_canceled',
    palette='viridis'
)
plt.title('Cancellation Rate by Lead Time')
plt.xlabel('Lead Time (Days)')
plt.ylabel('Cancellation Rate')
plt.gca().yaxis.set_major_formatter(plt.matplotlib.ticker.PercentFormatter(1.0))
plt.show()