Taxi
Taxi example, adapted from the NYC Taxi Rides example from Mosaic.
Pickup and dropoff points for 1M NYC taxi rides on Jan 1-3, 2010.
Click and hold to begin brushing in either map.
How does Pluot differ from Mosaic?
- Mosaic performs rasterization of the lon/lat coordinates and renders as a static image, while Pluot directly plots each coordinate and renders as a scatterplot allowing zoom and pan interactions
- Can you find Newark airport? Where are passengers who are dropped off there picked up, and vice-versa?
- The Mosaic example's histogram does not update in response to brushing in the map.
- Do rides to/from Brooklyn or Queens have the same temporal distribution as rides to/from Manhattan?
- Mosaic computes the histogram via the CPU using SQL and DuckDB, while Pluot computes the histogram via the GPU using a WebGPU compute shader.
- Mosaic uses the DuckDB spatial extension to perform the true map projection, while Pluot approximates the map projection using an affine transformation matrix.
- The Mosaic example filters on every brush interaction, while the Pluot brush waits until the interaction finishes to perform filtering.
- In the Pluot example, we color the points by the passenger count.
Open your browser network and memory tab, then reload each example page to check the memory and network overhead of each tool for yourself.