British Airways: Predictive Modeling of Customer Bookings

Project Overview

This project develops and interprets a machine learning model to predict the likelihood of a customer completing a flight booking. The goal is to support British Airways with actionable insights for smarter marketing, improved resource planning, and a better customer experience.

---

Dataset Description

The dataset (customer_booking (1).csv) contains detailed records of individual booking requests, including both customer and flight attributes.
Key features:

• num_passengers: Number of passengers per booking
• sales_channel: Booking method (Internet, Travel Agent, etc.)
• trip_type: Round trip or one-way
• purchase_lead: Days booked in advance
• length_of_stay: Length (days) for round trips
• flight_hour: Scheduled departure hour
• flight_day: Day of the week
• route: Route code (airport pair)
• booking_origin: Country of booking
• wants_extra_baggage, wants_preferred_seat, wants_in_flight_meals: Customer preferences
• flight_duration: Flight duration in hours
• booking_complete: Target variable (1 if booking completed, 0 otherwise)

---

Process Overview

1. Data Exploration & Preparation

• Loaded the dataset and explored variables to understand distributions, spot missing values, and check for duplicates.
• Categorical features (sales_channel, trip_type, flight_day, route, booking_origin) were converted to numeric using one-hot encoding.
• Checked and addressed missing values.
• Removed duplicates as needed.

2. Feature Engineering

• Created additional features likely to improve predictions, e.g., total seats, specific customer preferences, or timing bands.
• Retained engineered and relevant original variables for modeling.

3. Model Development

• Random Forest Classifier selected for strong performance and interpretability.
• Data split into training and test sets (80/20 split).
• Model trained to predict booking_complete.

4. Model Evaluation

• Model achieved approximately 86% accuracy on the test set.
• Additional metrics (precision, recall, F1-score) calculated to confirm robust performance.

5. Feature Importance Analysis

• Identified the top 10 features influencing bookings, such as:
  • Purchase lead (advance booking days)
  • Flight hour (departure time)
  • Length of stay
  • Number of passengers
  • Flight duration
  • Booking origin
  • Customer preferences (meals, seat selection)
• Visualized the top drivers using a bar chart.

---

Code Summary

Click to expand Python workflow

# 1. Data Loading & Exploration
import pandas as pd
df = pd.read_csv("customer_booking (1).csv", encoding="ISO-8859-1")
print(df.info())
print(df.head())

# 2. Data Cleaning
df = df.drop_duplicates()
df = df.fillna(df.median(numeric_only=True))

# 3. Feature Engineering & Encoding
cat_cols = ['sales_channel', 'trip_type', 'flight_day', 'route', 'booking_origin']
df_encoded = pd.get_dummies(df, columns=cat_cols, drop_first=True)

X = df_encoded.drop('booking_complete', axis=1)
y = df_encoded['booking_complete']

# 4. Train/Test Split
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# 5. Random Forest Model
from sklearn.ensemble import RandomForestClassifier
rf = RandomForestClassifier(n_estimators=100, random_state=42)
rf.fit(X_train, y_train)

# 6. Evaluation
from sklearn.metrics import accuracy_score, classification_report
y_pred = rf.predict(X_test)
print("Accuracy:", accuracy_score(y_test, y_pred))
print(classification_report(y_test, y_pred))

# 7. Feature Importance
import matplotlib.pyplot as plt
import numpy as np
importances = rf.feature_importances_
indices = np.argsort(importances)[-10:][::-1]
features = X.columns[indices]
plt.figure(figsize=(10,5))
plt.barh(features, importances[indices])
plt.xlabel('Importance')
plt.title('Top 10 Feature Importances')
plt.gca().invert_yaxis()
plt.show()