Car Dependency Index — interactive viewer

An interactive visualisation of the Car Dependency Index (CDI), a high-resolution measure that compares opportunity access by private car versus public transport for every ~200 m hexagonal cell of a city.

CDI is defined as

CDI = (O_car − O_PT) / (O_car + O_PT)

so it ranges from −1 (PT-favoured) through 0 (balanced) to +1 (car-dependent).

The viewer accompanies the paper Car Dependency in Urban Accessibility by Campanelli, Marzolla, Bruno, Melo & Loreto (2026 · arXiv:2604.01019) and is styled to match the companion accessibility-pov viewer for the upstream proximity/opportunity framework (Bruno et al., EPJ Data Science 2026).

Features

• World landing map — pins for every city in data/index.json. Cities with bundled hex-level data are highlighted in red; others are dimmed and carry a no data badge.
• City view — three-panel layout:
  • Cartogram (Leaflet base + Canvas hexagon overlay) with a continuous blue ↔ white ↔ red CDI gradient.
  • Sidebar with: scale legend, dual-thumb CDI range filter, live city summary, selected-hexagon inspector and interaction tips.
  • Scatter plot of opportunity-by-car (x) vs. opportunity-by-PT (y), with the y = x diagonal marking CDI = 0. Cross-highlights with the map.
  • Drag-to-resize handle between map and scatter; sidebar can be collapsed.
• Compare cities (Stats view) — diverging bar chart of population-weighted CDI, city-level scatter, population-weighted CDI distribution and a sortable summary table.
• About / map / scatter help modals explaining the methodology.
• Bilingual UI (EN / IT) toggled from the header.
• Mobile responsive layout with tabbed cartogram / scatter / info panels.

Data layout

data/
  index.json            ← city manifest (slug, name, center, zoom)
  <city-slug>/
    hexes.geojson       ← hexagonal polygons + properties.id
    cdi.csv             ← hexagon_id, o_score_pt, o_score_car, CDI, population, …

The viewer joins on properties.id (GeoJSON) ⇄ hexagon_id (CSV) at load time. To add a city, drop its folder under data/ and add an entry to data/index.json. The landing page automatically detects which cities have bundled data.

Reference

Campanelli B., Marzolla F., Bruno M., Melo H. P. M., Loreto V. (2026). Car Dependency in Urban Accessibility. arXiv:2604.01019

The companion accessibility framework on which this work builds is described in: Bruno M., Campanelli B., Melo H. P. M., Rossi Mori L., Loreto V. (2026). The dimensions of accessibility: proximity, opportunities, values. EPJ Data Science 15:22. doi:10.1140/epjds/s13688-026-00623-8