Fifteen million people take Uber rides every day, and they generate a lot of data.
In fact, two years ago Uber began building its own mapping tool to process all of it. Called Kepler.gl, the software maps location data and runs entirely in your web browser, processing up to a million data points to create beautiful map-based visualizations. Uber recently made Kepler.gl open source and available to the public, and now other companies–like Mapbox, Airbnb, Alphabet’s Sidewalk Labs, and scooter company Limebike–are all using it to work with their own data.
It’s no wonder: Typical data visualization mapping tools require a lot of know-how or long tutorials. But Kepler.gl’s interface is designed to be easy to use for anyone, even people without experience in cartography or data management. All you have to do is upload a CSV or GeoJSON file and it creates a map for you, no coding necessary. From there, you can apply filters that turn your data into different types of visualizations, like heatmaps, line maps, and arch maps, all of which are designed to help users discover patterns that would be hard to find without laying it all out visually.
Data viz engineer Shan He originally built Kepler.gl for Uber’s data science team. “They wanted a better way to show map data,” He says. “But because it’s web-based, we [saw] it picked up by [Uber’s] city operation teams, designers, the policy team, and also other engineers.”
So far, Uber’s teams have used the tool to look at the density of trips in certain areas so they can direct drivers to areas where there’s more demand, or to inform city officials that there are consistent traffic blockages in certain areas of a city. One team uses Kepler.gl to examine how drivers travel from part of the city to another, which is used to improve Uber Pool ride-share routing. “We allow engineers to plot many trips, connect them using the Arch layer, and look around to see what parts of the city are more interconnected with other parts of the city,” He says.
The policy team also uses the tool to demonstrate to city officials how popular Uber is by showing them heat maps of the company’s rides, and Uber says that both Airbnb and Alphabet’s Sidewalk Labs are using the tool for similar purposes, including data analysis and the presentation of data to cities.
After the tool opened to the public last month, Mapbox announced it was integrating Kepler.gl into its own browser-based mapping tool–and the company used it make a map of airplane flight patterns around Chicago’s O’Hare airport. Other individuals have taken to it as well. For instance, one person shared a map of bike-sharing trips in Hamburg, Germany. A student in public transport network design and optimization in Brazil made a map of daily bus boardings in São Paulo. Another person visualized their own Uber trips. He says people have used it to plot things like the refugee crisis, deforestation, and the state of coal mines.
Part of the appeal is that Kepler.gl works for any type of location-based data–and if you upload to the tool, everything stays in your browser. That means Uber doesn’t get its hands on whatever data you’re playing with. There are downsides to that as well: There is no easy way to share a map besides take a screenshot of it. Otherwise, you have to share a .json file and upload it to Kepler.gl on your browser.
Right now, He says the team is in the process of getting feedback from users and deciding what new features to build into the tool. You can try it yourself right here.