Navigating San Francisco's Hills: An Overview of Flatten SF

Created by Drew Edwards, Flatten SF is a specialized routing tool designed to help users identify the least strenuous paths between any two locations within the city of San Francisco.

🛠️ Core Functionality

The application allows users to input a starting point (From) and a destination (To). For cyclists, there is a specific toggle to Prefer calm streets, which adjusts the routing logic based on street safety and comfort.

The tool provides real-time data on:

• Total distance (Shortest)
• Total elevation gain (Flattest)
• Maximum steepness encountered

---

⚙️ How the Engine Works

Routes are calculated locally within your web browser. The system analyzes over 160,000 individual street segments, leveraging high-resolution USGS 1m lidar elevation data combined with the Overture / OpenStreetMap street graph.

The Trade-off Slider

The interface features a slider that manages the tension between distance and vertical climb. It identifies the "Pareto frontier"—routes where no other path is simultaneously shorter and flatter.

The Logic:

• Moving the slider right: The route will either stay the same length or get longer, but it will never increase in climbing.
• The Limit: The flattest possible route is determined by a specific cost ratio: $\text{1 ft of climbing} = 200 \text{ ft of walking}$ If a detour exceeds this ratio, the system stops considering it a viable route.

Understanding "Climbing"

It is important to note that the tool measures cumulative gain. Net elevation difference between start and end $\rightarrow$ Total ascent.

---

🚲 Cycling & "Comfort" Metrics

When the "Prefer calm streets" option is enabled, the algorithm stops measuring raw distance and starts measuring "comfort distance." This is calculated by applying a multiplier to the actual length of the street segment:

// Conceptual logic for comfort distance
function calculateComfortDistance(length, streetType) {
  const weights = {
    'protected': 0.8,
    'quiet': 1.0,
    'arterial': 1.7 // average of 1.4 - 2.0
  };
  return length * weights[streetType];
}

---

🗺️ Data Architecture

The search functionality is entirely offline. All necessary data—including street intersections, specific addresses, and points of interest within San Francisco—are embedded directly into the page.