Uber has opened source its internal visualization toolkit, Ketoper.gl, which is a high performance React component built on Deck.gl for visual exploration of large geolocation data sets. Its support for GPU capabilities allows applications to render millions of data points at once.
Uber mentioned that to visualize data, regardless of which development framework is used, developers need to invest a lot of time and resources, the process is always boring, and it can’t realize all the ideas of data visualization. And Kepler.gl is a useful way to solve that problem.
Kepler.gl supports CSV, JSON and GeoJSON formats. Users can drag and drop these files with a simple click. In the Web interface, you can browse through this data by manipulating different layers, performing filtering or aggregation, and so on. Finally, you can choose to export the map statically and dynamically. Uber says processing data that used to take weeks can now be done in 10 minutes.
All layers are gPU-accelerated due to support for WebGL development mechanism deck.gl. Even for Web applications, Kepler.gl can successfully map millions of data points, meaning it can render thousands of trips over and over, as well as do Spatial Aggregations.
In addition to offering traditional two-dimensional X – and Y-axis plane graphics, Kepler.gl also supports the addition of z-axis solid images, allowing users to draw grids, hexagons or polygonal cylinders, thereby increasing the amount of information a map can convey.
Kepler.gl can also conditionally filter data, and users can use any data indicator to remove or filter out data that meets the criteria, typically by creating a visual time data with a timeline.
