Godwit - The Extended Airline Pathfinder

Problem Statement

The airline booking service and other OTAs generally provide only direct or low-layover flight. This creates the demand of time-generous price-conscious flyer underserviced by the traditional booking interface. Godwit intends to provide improved availability through highly customizable airway pathfinding for the such cohorts.

Hypothesis

By utilizing the pathfinding with relaxed restriction of the airline network as graph, the pathfinding will offer more flexible choice such that more affordable and time-extended path, or intentionally flying through less season-based demand to improve price predictability that existing vendor do not currently offer.

Proposed Approach

Given the constraint that acquiring real time fare data that is volatile, and expensive to access, we are considering to utilize A* algorithm with the posted fare as heuristics to determine candidate path under regular fares. Data acquisition remains the primary open question. The pathfinding approach is designed to be data-source agnostic, with evaluation ongoing across public fare sources, BTS datasets, and middleware providers such as Duffel.

Open Question

• Do seasonal or demand based fare change meaningfully differentiate the optimal path, or the optimal path remains to be consistent throughout the year?
• Does round trip discount consideration change pathway significantly?

Out of Scope at Current State

At the current state, we aim to proof the concept of this hypothesis using various available data, not providing the service that comparable to the other OTA services, therefore we will exclude the following from the initial functionarity:

• Self-connecting itinerary connection risk
• International routing / visa acceptability consideration
• Layover / Lounge information

Next Step

We intend to create proof of concept using currently available data while attempt to acquire more realistic data source from possible vendor.