FLIGHT//DB
Interactive analytics across 104 airports and 532 routes: delay patterns, revenue concentration, fleet utilization and the network geography underneath them.
POSTGRESQL 16BIGQUERYPYTHON ETLNEXT.JS
0
Records analyzed
bookings, tickets, flights and boarding passes
0
Airports
spanning 11 time zones
0
Total revenue (RUB)
across 451 revenue-generating routes
0.0%
Avg delay rate
flights departing >15 min late
0x
Mat view speedup
same query, materialized view vs raw
0
ETL migration (sec)
PostgreSQL to BigQuery, end to end
0
Flights arrived
the basis for every delay figure
0
Routes mapped
between 104 airports
EXPLORE THE DATA
Six views, one network. Follow them in order like chapters, or jump straight to what interests you.
ROUTE MAP
01
Where is the network dense, and where does it hang by a single arc?
104 airports, 532 routes on an interactive WebGL map
DELAY ANALYSIS
02
When do flights run late, and do they ever catch up in the air?
Heatmaps, worst routes, aircraft and time-of-day patterns
REVENUE
03
How few routes carry half the money?
Pareto curves, fare class split, monthly trends
FLEET
04
Which aircraft earn their keep, and which fly nearly empty?
Load factors, turnaround times, seat configurations
PERFORMANCE
05
How does a 381 ms query become 0.13 ms?
Query plans, index strategy, partitioning, storage reclamation
DATA PIPELINE
06
When does BigQuery beat PostgreSQL, and when is it 200x slower?
Warehouse migration + engine performance comparison