Flight Booking and Customer Reviews

Project Overview

Worked on a simulation demonstrating the critical role of data science in British Airways' success. The project involves scraping and analyzing customer review data to uncover key insights and building a predictive model to identify factors influencing buying behaviour.

Name	Description
Review	Customers' comments on specific airlines flown
Rating	Overall rating on specific airlines flown
Country	Customers' country of origin
Type Of Traveller	Type of customer (e.g., Business Traveler, Leisure Traveler, etc.)
Seat Type	Class of the seat (e.g., Economy, Business, First Class)
Date Flown	Date when the customer flew with the airline
Recommended	Indicates if the customer recommends the airline (Yes/No)
Aircraft	Type or model of the aircraft flown
Seat Comfort	Customer's rating of seat comfort
Cabin Staff Service	Customer's rating of the service provided by cabin staff
Food & Beverages	Customer's rating of food and beverages
Inflight Entertainment	Customer's rating of inflight entertainment options
Ground Service	Customer's rating of ground services provided by the airline
Wifi & Connectivity	Customer's rating of the onboard wifi and connectivity
Value For Money	Customer's rating of the overall value for money
Month Flown	Month when the flight was taken
Year Flown	Year when the flight was taken
Departures	Departure airport or city
Arrivals	Arrival airport or city
Layover	Indicates if there was a layover and its details (if applicable)

Project Objectives

1. Data Cleaning & Preprocessing: Handling missing values, normalizing text, and feature cleaning.
2. Exploratory Data Analysis (EDA): Understanding distribution patterns, trends, and relationships.
3. Natural Language Processing (NLP): Sentiment analysis using VADER and word cloud analysis.
4. Predictive Modeling: Implementing classification models to predict customer recommendations.
5. Model Evaluation & Interpretation: Assessing model accuracy and analyzing feature importance.

Machine Learning Models Used

• Natural Language Processing (NLP): VADER sentiment analysis, word cloud visualization.
• Random Forest Classifier: Used for predicting customer recommendations.
• Feature Importance Analysis: Identifying key factors that influence customer recommendations.
• Evaluation Metrics: Accuracy score, confusion matrix, classification report.

Technologies Used

• Programming Language: Python
• Libraries: pandas, numpy, scikit-learn, regex, matplotlib
• Visualization Tools: plotly, seaborn

Project Workflow

1. Data Collection: Gather customer reviews and flight booking data.
2. Data Cleaning: Process missing values, normalize text data, and clean categorical fields.
3. Exploratory Data Analysis: Generate summary statistics, visualizations, and sentiment insights.
4. Feature Engineering: Extract relevant attributes for better prediction accuracy.
5. Model Training: Train a Random Forest Classifier for recommendation prediction.
6. Model Evaluation: Analyze accuracy scores, classification reports, and confusion matrices.
7. Results Interpretation: Identify key insights from feature importance and sentiment analysis.