Air Canada No Show Rate Analysis
Project Overview
This project aims to analyze Air Canada's passenger no-show rates by combining real-time flight data with external factors that influence passenger behavior. The analysis uses a comprehensive data collection pipeline to gather multiple data sources and simulate passenger booking patterns.
What Was Done
1. Real-Time Flight Data Collection
Uses the OpenSky API to collect current Air Canada flights
The AirCanadaDataCollector class filters flights by callsign starting with "ACA"
Collects flight attributes like position, altitude, velocity, and operational status
2. Route Analysis
The AirCanadaRoute class identifies flight patterns across major Canadian hubs
Focuses on three major hubs: Montreal (YUL), Toronto (YYZ), and Vancouver (YVR)
Analyzes route frequency and hub traffic distribution
3. External Data Integration
The ExternalDataCollector gathers weather data from Canadian weather APIs
Tracks Canadian holidays and seasonal factors
Simulates realistic weather conditions when API data is unavailable
4. Passenger Booking Simulation
The PassengerBookingSimulator generates realistic booking patterns
Factors affecting no-show probability:
Advance purchase timing (1-90 days)
Fare class (Basic, Standard, Flex, Premium, Business)
Passenger type (Business, Leisure, VFR)
Uses research-based assumptions (12% baseline no-show rate)
5. Integrated Data Pipeline
The AirCanadaPipeline class orchestrates the entire data collection process
Combines flight data, route analysis, external factors, and passenger simulations
Produces comprehensive flight analysis with predicted no-show rates
How This Is Useful
This analysis could be used for:
Revenue optimization through better overbooking strategies
Operational planning by predicting passenger loads
Route performance analysis across different hubs
Seasonal demand forecasting
The project demonstrates a sophisticated approach to aviation data analysis by combining real-time operational data with passenger behavior modeling.
