Research Summary: A global mobile network coverage raster product at 1km resolution, 1999--2030

A new global, sub-national dataset has been developed, providing annual 1km resolution maps of 2G, 3G, and 4G mobile network coverage probabilities for 214 countries and territories from 1999 to 2030. This record is crucial given the impact of mobile signal availability on various aspects of societal and economic activity in the digital age. The dataset combines outputs from three independent models: a machine-learning model, a techno-economic simulator, and a spatial deep-learning model, offering a single best estimate with uncertainty bands.

Why it matters

This initiative provides an unprecedented, granular view of global mobile network infrastructure, which is foundational for understanding digital inclusion and economic development. The data can inform strategic investments, policy-making, and operational planning across various sectors dependent on mobile connectivity, ensuring resources are allocated effectively where they are most needed.

Key insights

• A globally consistent, sub-national record of mobile network coverage (2G, 3G, 4G) now exists, addressing a prior gap in data.
• The dataset spans 1999 to 2030, offering historical data and future projections for 214 countries and territories.
• Coverage is mapped at a 1km resolution, enabling granular analysis of mobile signal availability.
• The product integrates insights from three distinct models (calibrated machine-learning, techno-economic simulator, spatial deep-learning) to generate robust estimates.
• Each estimate includes per-pixel 90% uncertainty bands, enhancing data reliability and utility.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2609.11320