FareIQ - Team 25

Feature Importance
Model Capabilities
result metrics models
model graph

Inspiration

Airfare pricing feels random to consumers, but it’s structural! We were inspired to quantify how competition, hub dominance, and low-cost carrier presence systematically shape fare outcomes to help others and stakeholder give actionable insights.

What it does

FareIQ is an interactive analytics platform that identifies when and where airline routes are structurally over or under priced, and highlights markets vulnerable to fare spikes due to weak competition.

Our visualization turns that model into an actionable tool:

Route thickness = passenger demand
Route color = pricing deviation (blue = underpriced, red = overpriced)
Time slider = watch pricing dynamics evolve across 14 quarters

Users can instantly identify:

High-traffic routes charging above structural expectation
Markets vulnerable to fare spikes if LCC competition disappears
Routes where pricing power exceeds competitive fundamentals

How we built it

We analyzed 14 quarters of U.S. DOT route-level data (2022–2025), 4 mew engineered features and using rolling validation to fit 6 models (Linear Regression, Ridge, XGBoost, HGB, Random Forest, Extra Trees) to predict expected fare price. Finally, we visualized pricing deviations on an interactive U.S. route map using react.

Challenges we ran into

Causal inference around airline exits was noisy, quarterly data limited seasonality granularity, and avoiding data leakage in pricing models required strict time based validation. We also ran out of time since this is one day hackathon so lots of time restriction. We also didn't have much time to price our pitch so it is definitively something to improve on.