SPARC Building Scan

Welcome to the SPARC Building Scan page. This website provides an introduction to the Building Scan data product, serves as a guide for technical users, and documents the detailed methodology behind the analysis.

The Building Scan product provides detailed insights into a city’s building stock, including building types, heights, footprints, and other relevant attributes. This enables city planners, policymakers, and investors to make informed decisions regarding urban development, infrastructure planning, and risk assessment.

Building Scan Workflow

The Building Scan product generates high-quality, analysis-ready building footprints for a given city. Building footprints are digitized polygons outlining each building across an area of interest. They are a foundational data layer for urban planning, climate risk and resilience analytics, infrastructure and service gap assessments, and cadastral systems.

Project teams often default to public sources such as OpenStreetMap or Google Open Buildings. While these are valuable resources, their accuracy varies greatly by location. Their buildings may be outdated or missing in informal and recently developed areas, contain faulty polygons, or lack key attributes. Building Scan provides a consolidated workflow to help teams choose the most suitable building layer from three options.

1. Open-Source Layers

Integrates public datasets (Overture, OSM, 3D GlobFP), with cleaning to remove duplicates and false positives.

Typical use: Investment scoping.

2. AI Extraction

Option 1, supplemented with AI extraction in poorly mapped areas (missing or outdated buildings), using high-resolution satellite imagery and the open-source Segment Anything Model (SAM3).

Typical use: Analytics informing infrastructure operations.

3. Manual Digitization

Digitization carried out by a specialist firm using procured imagery.

Typical use: Taxation, or a complete building registry or cadastral system.

Comparison of building scan workflows
Option 1 — Open-Source Option 2 — AI Extraction Option 3 — Digitization
Approach Integration of public datasets Option 1 plus AI-derived buildings from imagery Digitization by a specialist firm using procured imagery
Output Footprints with height and year, plus completeness assessment High-coverage footprints with height and year Full-coverage footprints with height and year
Cost < $5k ~$10k ~30k
Turnaround 2 weeks 4 to 6 weeks 3–6 months
Accuracy Moderate High (~0.8) Very high (~0.9)

Data Attributes

Each building is assigned the following attributes.

  • Building Heights: Information on the vertical dimension of buildings, which is crucial for urban planning and risk assessment.
  • Building Age: Information on the age of buildings, which can be important for understanding the historical development of the city and for assessing the need for maintenance or retrofitting.
Map of building outlines across Juba
Figure 1: Building footprints by year, Juba
Map of building outlines across Juba
Figure 2: Building footprints by height, Juba

SPARC Project Background

Cities in developing countries often lack current, reliable, and accessible geospatial data. Existing information is frequently fragmented across analog or disconnected systems, costly to acquire and analyze, and difficult to reuse because of inconsistent licensing and limited technical capacity. The Smart Parcel Atlas for Resilient Cities (SPARC) program addresses this gap by creating a shared, scalable platform for urban geospatial intelligence. There are three tiers of the SPARC program: 1. Tier 1 — Global urban data and analytics 2. Tier 2 — City-specific geospatial intelligence 3. Tier 3 — Operational uptake and client integration

Tier 1 — Global urban data and analytics

Provides standardized, quality-assured geospatial data and automated statistics for approximately 10,000 cities worldwide. Supports early operational scoping and serves as a “conversation starter” for country and city engagements. Made available through the SPARC Toolkit on GeoInsights for internal and external users.

Tier 2 — City-specific geospatial intelligence

Delivers higher-resolution, on-demand analytics for up to 100 priority cities. Supports investment option assessment and site prioritization through six initial services: buildings and parcels; urban public infrastructure; extreme heat exposure; multi-hazard exposure; green infrastructure; and urban settlement forecasts.

Tier 3 — Operational uptake and client integration

Connects SPARC analytics directly to World Bank Group operations and client decision-making. Regional operational teams lead demand identification, client engagement, local validation, and incorporation of findings into projects and local systems.