How Uber’s Data Visualization Team Turns Billions of GPS Points into Interactive Maps

Overview

In early 2015 Uber formed an official data‑science team to discover insights through visual data‑exploration tools. Managing billions of GPS locations and millions of events per minute, the team built visualizations that let stakeholders instantly see missed business opportunities without writing code.

The team grew from two engineers to a 15‑person full‑stack group covering computer graphics, information design, creative technology, and web development. Their focus spans visual analysis, map rendering, framework development, and public‑facing data storytelling.

Visual Analysis: Enhancing Data Operability

Visual analysis at Uber mainly involves abstract data visualizations without inherent geographic structure, contrasted with scientific visualizations that map physical worlds. Most effective visual analyses are dashboards, real‑time charts, and network graphs that support reporting and exploration.

The team emphasizes reusable components and recently open‑sourced react-vis, a React‑based visualization library built on an enhanced D3, offering JSX‑style syntax for declaring axes, chart types, and other visual elements.

Map Rendering: Big‑Data Exploration

Uber’s GPS data is its richest asset, with billions of points collected daily. Rendering this data intensively in browsers posed major challenges for real‑time map visualizations.

Different user groups—city managers, operations teams, and data‑science analysts—receive custom map applications that provide live supply‑demand distributions, aggregated market insights, and multidimensional drill‑downs.

The technical stack relies on open‑source libraries the team previously built: react-map-gl (a React wrapper around Mapbox GL) and deck.gl, which creates WebGL‑accelerated layers that can overlay or replace map content.

Public‑Facing: Telling Data Stories

Uber uses data visualizations to communicate stories about safety, efficiency, traffic, and policy to a broad audience. One recent project visualized how Uber POOL reduces city traffic, showing side‑by‑side maps of congestion with and without the service.

Collaborating with design teams, the group also built a 3D animated map that animates every Uber trip of a day. The pipeline pulls data from Hive, renders frames server‑side with WebGL, and compiles them into video.

For these high‑performance visualizations the team created the luma.gl framework, a WebGL‑focused library built on modern ES6, WebGL 2.0, and component‑based architecture, interoperable with other libraries such as stack.gl.