Improve the Passenger Load Factor of Australian Domestic Airline to 82.5% by arriving at an optimum number of seats from Adelaide to six cities - Brisbane, Canberra, Gold Coast, Melbourne, Perth and Sydney for time period September 2016- September 2017.
Passenger Load Factor – Passenger load factor measures the capacity utilization for airlines. It signifies the efficiency with which an airline fills seats and generates revenues. 80% of passenger load factor is considered as standard in the domestic airline industry.
Build a model to forecast the aggregate passenger traffic on Australian Domestic Airlines for each month from September 2016 – September 2017 based on prior values. Evaluate multiple models and identify the optimum model based on MAPE (Mean Absolute Percentage Error), goodness of fit, distribution of residuals and auto correlation of residuals.
Data Set Overview
Monthly data with the respective passenger load factor for each of the 6 cities is available.
- Data Set- Airline passengers to 6 cities
- Source- Kaggle (https://www.kaggle.com/alphajuliet/au-dom-traffic)
- Training data – 42- 66 months of data
- Validation data - 18 months of data set
- Forecasting horizon - forecast passenger trips for 12 months Sep 2016- Sep 2017
Forecasting Model Overview
The data set was split into 6 different time series based on various cities – Canberra, Sydney, Melbourne, Perth, Gold Coast, and Brisbane. For each of the time series various forecasting models were based on the specified criteria and the optimum model was identified. (refer section Forecast Methods for more information)
- Based on the forecasted seats calculate the maximum capacity needed and the number of flights based on below equations
- Max Seats = Forecasted Traffic / 82.5%
- Number of Flights = Max Seats/ 50 (1 domestic flight = 50 seats (appx))
- Based on the forecasted number determine the fleet optimization plan
- Flight Planning and Schedule Development
- Maintenance schedule
- Flight lease plan
As the forecasted period is 6 months ahead the airline has enough time to optimize the number/type of planes on the route. Determining optimal number of flights would help Australian airlines in better scheduling of its flights and effective capacity utilization. Decisions such as leasing new flights or scheduling maintenance plans for individual aircrafts can be done effectively if an estimated demand is available with the airlines. The estimated demand will hold true given underlying conditions remain unchanged. This model would work well if the forecasts are updated over time based on the actual values.