Neighborhood Experience Discovery

This project is an early-stage prototype and part of an ongoing investigation into walkability, retail ecosystems, and everyday urban decision-making. Using Central Square, Cambridge as a case study, this tool combines walkable isochrones, local retail data, and free-text search to support experience-based discovery, rather than destination-based navigation.

What kinds of local experiences are discoverable within a short walk, and how does walk time shape what people can meaningfully do in a neighborhood?

Interactive Demo: discover neighborhood experience

Use the tool to

  • Select walk time

  • Search by experience or need

  • Explore store details and build a walkable itinerary

Use the Walk Time selector in the top-right corner to switch between:

  • 5 minutes

  • 10 minutes

  • 15 minutes

Each option visualizes the area reachable on foot within that time window.

Use the search bar to explore the neighborhood using natural language, such as:

  • “gifts”

  • “local”

  • “second hand”

  • “bakery”

Results update dynamically and highlight stores whose descriptions match your query.

Click on a store marker to see:

  • Store name

  • A short description snippet

  • A direct link to the store’s website

  • An option to add the store to your itinerary

Methodology

Data

  • Central Square BID business dataset

  • OpenStreetMap street network

Walkable Isochrone Modeling

  • Built a pedestrian street network using OSMnx

  • Calculated walking travel times assuming a consistent walking speed

  • Generated 5, 10, and 15-minute walking isochrones from the neighborhood center

  • Exported isochrones as GeoJSON layers for use in the interactive map

Interactive Map Development

  • The front-end interface was built using Leaflet.js, with custom controls layered on top:

    1. Toggleable isochrone layers (5 / 10 / 15 minutes)

    2. Free-text search across local businesses

    3. Interactive store popups

    4. Lightweight itinerary builder

Limitations

Data

  • Website text scraping is imperfect and varies widely by store

  • Limited or noisy online content affects text-search quality

  • Retail data reflects a snapshot in time and does not capture openings, closures, or temporal rhythms

Modeling

  • Isochrones do not account for elevation, signal delays, or sidewalk quality

Feature Scope

  • Search is keyword-based rather than semantic

  • Routes are not auto-optimized and itinerary planning is manual

  • The tool does not yet adapt to user context (time of day, weather, mobility needs)

Previous

Prediction: Retail Vitality