Skip to main content
Uber logo

Schedule rides in advance

Reserve a rideReserve a ride

Schedule rides in advance

Reserve a rideReserve a ride
Data / ML, Engineering

Visualizing Traffic Safety with Uber Movement Data and Kepler.gl

May 17, 2019 / Global
Featured image for Visualizing Traffic Safety with Uber Movement Data and Kepler.gl
Figure 1: Kepler.gl, an open source data visualization tool, can be installed locally or used online as a web-based application.
sa.geojsonfreeflow_speedhourspeed_difference
{“type”:”LineString”,”coordinates”:[[-73.9907278,40.7559486],[-73.9878881,40.7547554]]}11.80124223602484422:006.211180124223602
{“type”:”LineString”,”coordinates”:[[-73.9947719,40.7699966],[-73.9948449,40.7697672],[-73.9949029,40.7696256],[-73.9950615,40.7693328]]}32.9192546583850919:006.832298136645963
Figure 2: The Kepler.gl web-based application reads datasets from uploaded files and URLs, and also offers previously loaded sample datasets to help users get started.
Figure 3: After loading our Uber Movement dataset from New York City, we see a map of Manhattan. However, as we have not specified which specific data we want to visualize, there are no insights to gather from this visualization.
Figure 4: Applying a color to the speed_differences data in Kepler.gl now shows us where drivers tend to go faster in Manhattan.
Figure 5: Once we add a filter and specify the hour field from our dataset, Kepler.gl applies a playback control in its interface.
Figure 6: Kepler.gl creates a visualization of three datasets, letting us see how data collected from different organizations intersects.
Figure 7: When we zoom in on the intersection of Canal and Bowery Streets, we can see a number of warning signs from our data, from vehicle speed to fatal crashes.
the_geomOBJECTIDLMNOName
MULTIPOLYGON (((-122.440573525 37.731716381, -122.440711839 37.731715649, -122.440713234 37.731934272, -122.440574917 37.731934574, -122.440573525 37.731716381)))15478Sunnyside Conservatory
MULTIPOLYGON (((-122.427201866 37.763824121, -122.4272296 37.764103529, -122.426541611 37.764142947, -122.42652189 37.763937175, -122.426825127 37.763886758, -122.426881817 37.763877333, -122.427201866 37.763824121)))1691Mission Dolores
MULTIPOLYGON (((-122.454960515 37.771900069, -122.454830759 37.771960595, -122.454774695 37.771884863, -122.454568065 37.771981247, -122.454463754 37.771840342, -122.454800139 37.771683432, -122.454960515 37.771900069)))276175McLaren Lodge
Figure 8: Kepler.gl creates a visualization of the dataset, in this case showing landmarks included in the data on a map of San Francisco.
Figure 9: Kepler.gl lets us customize colors and fills to make our data visualization more understandable.
Carsten Jacobsen

Carsten Jacobsen

Carsten Jacobsen is an open source developer advocate at Uber.

Posted by Carsten Jacobsen