Applying Mapping and Spatial Analytics: Volume 2 by Dawn J. Wright & Christian Harder

Author:Dawn J. Wright & Christian Harder [Wright, Dawn J. & Harder, Christian]
Language: eng
Format: epub
Published: 2020-12-15T00:00:00+00:00

Neighborhood mapping workflows applied to satellite imagery in Caracas, Venezuela, identify building categories to generate insights on the socioeconomic patterns of urban land use.

Getting a good fit

Given the scope of the ORNL program, which aims to map every pixel across the globe, the organization will need more than one model to meet the challenge. Ideally, the fewer the models there are, the better. Current research is testing a proof-of-concept workflow with multiple models that characterize sample imagery across the planet. The concept is based on observations that pockets of similarities tend to reappear globally that are not necessarily tied to regional geographies. The idea is to analyze the entire collection of global imagery using an algorithm that can gather similar pixels and distribute them into multiple “buckets” for unique models. When unseen imagery enters the workflow, the computer quickly determines which bucket to use and pushes the input through the appropriate model. The beauty of the highly automated workflow is that once the buckets are allocated, the corresponding models can be changed as needed to support not only neighborhood mapping but also other kinds of analysis.

Estimating global populations—a people-per-pixel approach

To populate settlement and contextual layers at scale with the critical “missing pixels”—people—researchers must overcome global disparities in data availability. While the United States conducts regular censuses and household surveys that provide population insights at the national level, many other countries do not record adequate, up-to-date, or reliable population information (if any at all) at the minimum coverage needed to estimate and distribute populations with any degree of confidence. That means outside of North America, top-down disaggregation methods of population modeling are impractical. To scale population modeling for the world, bottom-up solutions for aggregating populations in any country are essential.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.