GM Safety View V2

Quick Highlights:

• 42% faster risk identification across pilot programs

• 5 state DOTs onboarded

• Built with the GM Aurora Design System

Problem Framing

User Pain Points:

• Difficult to find crash-prone areas from raw datasets

• Lack of consistency across existing mapping tools

• Overwhelming visual complexity in traditional GIS dashboards

• Limited ability to correlate events like near-misses, weather, and time of day

Primary Users:

• Department of Transportation safety planners and analysts

• Policy makers responsible for infrastructure prioritization

Key Question:
How can we turn connected vehicle data into a visual language that drives safer roadway decisions?

Research Insights

Approach:

• Conducted interviews with DOT analysts and traffic engineers

• Audited existing tools used by Vision Zero and other safety programs

• Mapped out real-world workflows for crash reporting and road safety auditing

Findings:

• Planners faced an overwhelming volume of unstructured data, where credibility and source reliability were often questioned.

• There was a lack of predictive analytics capable of supporting proactive safety measures rather than reactive reporting.

• Staffing and funding limitations across agencies slowed the adoption of innovative safety initiatives.

• Data collection standards varied widely between states, making it difficult to establish unified benchmarks.

• Urban counties were better equipped with modern GIS tools compared to rural counterparts, creating a gap in accessibility and data visualization capabilities.

Before Snapshot:
Legacy dashboards displayed raw CSV tables and low-resolution maps, requiring analysts to manually interpret trends.

Ideation and Design Process

Information Architecture:
Restructured navigation around user intent — Identify Risks, Explore Corridors, and Generate Reports — instead of data sources.

Wireframes and Explorations:

• Designed modular cards for key insights

• Created a contextual filtering system that updates visual layers dynamically

• Tested different visualization hierarchies for map density and overlay clarity

Visual System:

• Adopted GM Aurora design tokens for color, grid, and typography

• Designed high-contrast data visuals optimized for map overlays

• Implemented adaptive layouts for desktop and large display environments used in control centers

Key Decisions:

• Replaced dense charts with progressive disclosure of data layers

• Introduced “Insight Layers” toggles for weather, traffic density, and near-miss data

• Unified interaction patterns for filtering, exporting, and collaboration

Final Design and Experience

The Dashboard:

• Interactive U.S. map showing crash likelihood zones, near-miss patterns, and weather correlations

• Comparative corridor analysis feature for safety investments

• Configurable overlays and contextual tooltips for quick insight delivery

• Seamless export to CSV, PDF, and API for government reporting