In March 2018, a Data Scientist and a Full Stack Developer from lloopp presented this study as an entry to Accenture’s 2018 Data Challenge, where lloopp placed First Runner-up for the Best Predictive Model Category.
Can an airline save billions of pesos in cost of operations by unlocking the power of your data?
lloopp’s team took on this challenge and discovered that an airline can vastly save on operations cost by improving its Forecast Accuracy when scheduling its flight trips.
Using public data from Open Data Philippines (Data.gov.ph), lloopp made a study based on historical data from the Civil Aviation Authority of the Philippines from 2001 to 2006. Data was divided into three segments: Aircraft trips, Cargo freight trips and Passenger trips.
Using a Predictive Model Approach, lloopp performed a Time Series Analysis on all three data segments to see what relationships and patterns can be gleaned from trips made from 2001 to 2006. From this correlation, lloopp has discovered that the number of passengers highly influence the number of flights scheduled.
This generally means an airline’s operation costs depend heavily on passenger trips, more than cargo freight trips. By utilizing a good predictor for passenger movement, an airline would be able to effectively regulate aircraft trips monthly to minimize operations cost and maximize plane seating utilization.
Based on lloopp’s analysis, they developed a formula that can improve accuracy for Passenger Movement Forecasting. In short, if you accurately predict when passengers will be flying, you save on operations costs when at least 90% of the seats are booked. In this case, when this formula was applied to the available data, lloopp discovered that an airline can save up to Php20 billion in annual operation costs just through Passenger Movement Forecasting.