Since most interactions only involve a single dimension, and then only small adjustments are made to the filter values, incremental filtering and reducing is significantly faster than starting from scratch. Crossfilter uses sorted indexes (and a few bit-twiddling hacks) to make this possible, dramatically increasing the performance of live histograms and top-K lists. For more details on how Crossfilter works, see the API reference.
The coordinated visualizations below (built with D3) show nearly a quarter-million flights from early 2001: part of the ASA Data Expo dataset. The dataset is 5.3MB, so it might take a few seconds to download. Click and drag on any chart to filter by the associated dimension. The table beneath shows the eighty most recent flights that match the current filters; these are the details on demand, anecdotal evidence you can use to weigh different hypotheses.
Some questions to consider: How does time-of-day correlate with arrival delay? Are longer or shorter flights more likely to arrive early? What happened on January 12? How do flight patterns differ between weekends and weekdays, or mornings and nights? Fork this example and try your own data!