Trend Analysis

Turbulence Ahead: Ten Years of U.S. Flight Patterns

FAA and DOT data (2009–2018) visualize how delays propagate across the system, what carriers do to fight cancellations, and how weather or late aircraft shape the network.

Sampling Disclaimer: This visualization utilizes a dataset that was significantly sampled to adhere to Tableau Public's 150,000 row limit. Consequently, this is a modified representation of the original work, and specific visual trends may differ from the initial full-scale analysis.

Executive Summary

The Predictability of Chaos

While delays seem random, they are highly predictable. Fluctuations typically stay within 10-30% of flight volume, and forecast models show high confidence. The analysis confirms that seasonal delay rates consistently stay within 1 standard deviation of their historical averages.

Design Approach

Confidence Bands (LOD): Dotted reference lines indicate seasonal norms (1 std dev). Connected Trails: Airline performance is visualized as trails flowing from bottom (2009) to top (2018), with better performers (low delay) floating to the top.

Key Insights

Time of Day: Delays spike at the start and end of the day. Mid-year (summer) exacerbates these late-day operational fatigues.
Traffic ≠ Delay: High-traffic airports do not correlate with higher delay rates; volume is managed effectively.
Airline Strategy: Cancellation is a deliberate choice, not just a result of delay. Delta and Spirit show unique improvement trends over the decade.

The Dataset

BTS On-Time Performance (2009–2018). Limitations: Covers domestic non-stop flights only. Excludes small regional carriers (<0.5% revenue). Treats all flights equally (a 50-seat jet counts the same as a 300-seat widebody), meaning passenger impact is approximated.

Visual Strategy

Heatmaps reveal the "dark" start and end of operational days. Connected Scatterplots show the 10-year trajectory of airlines. LOD Calculations normalize seasonality to show true performance deviations.

Inference

Cancellation rates and delay rates are loosely correlated, suggesting that whether a flight gets cancelled is often a strategic decision to protect the network, rather than a direct result of delay duration.