median25th–75th percentile· fare index: 100 = the typical price for that trip; each itinerary is indexed to its own median before pooling, so sources and seasons can be compared
Long-run seasonality: what people actually paid (DOT, by quarter)
Q1Q2 / Q4Q3· passenger-weighted average one-way market fare actually paid (this direction), US DOT DB1B 10% ticket sample
Method
Every night the collector pulls the cheapest fare per departure date for the next 11 months from Aviasales search data (Travelpayouts), and, on a rotating schedule, runs a Google Flights search for US routes that also returns Google's 60-day price history for that itinerary. The heatmap and calendar use the latest sweep; "good" and "expensive" are the bottom and top quartile of that sweep. The booking curve pools every observation ever made for this route by days between observation and departure. DOT DB1B is the US Department of Transportation's 10% sample of tickets sold, aggregated by quarter. Prices are indicative, not bookable.