Analytic Infrastructure

Building robust, reproducible data infrastructure to bridge the gap between climate science and health research.

The Challenge

Climate and environmental data are abundant, but they are rarely structured for health research use. Data from weather stations, satellites, and reanalysis models arrive in different formats, units, temporal resolutions, and spatial coverages. Health researchers who want to study climate-health linkages face a steep technical barrier just to assemble a usable dataset — before any analysis can begin.

CGHI is building the infrastructure to solve this problem systematically. Rather than each research team independently wrestling with climate data access and processing, CGHI provides shared tools, pipelines, and curated datasets that make climate data accessible and practically useful for health researchers across the observatory network.

Data infrastructure and server systems — CGHI builds shared infrastructure so research teams can focus on analysis, not data wrangling.

Data Pipeline

Inventory & Access

Systematic cataloging of available climate, environmental, and health data sources across observatory regions — including weather stations, satellite products, reanalysis datasets, and health surveillance systems.

Harmonization

Standardizing units, temporal resolution, spatial alignment, and metadata across disparate data sources to create analysis-ready datasets. This is where tools like weathertools and dataplane are critical.

Quality Control

Automated and manual quality assurance processes to identify missing data, instrument errors, unit inconsistencies, and other data quality issues before analysis.

Documentation & Metadata

Embedding comprehensive metadata — variable definitions, units, provenance, and processing history — directly into datasets so they remain self-describing across projects, teams, and time.

Analysis & Dissemination

Reproducible analytic workflows and shareable outputs that enable cross-site comparison, synthesis, and collaborative research across the observatory network.

Weather station on a grassy mountain peak — Environmental monitoring stations provide the raw observations that feed CGHI's data pipelines.

Design Principles

Reproducible

Every step from raw data to analysis-ready dataset is documented, scripted, and version-controlled. Results can be independently reproduced.

Self-Describing

Datasets carry their own documentation — variable definitions, units, provenance — so they remain interpretable without external context.

Researcher-Facing

Infrastructure is designed for researchers, not just data engineers. Tools are documented, well-tested, and designed to minimize the climate data learning curve.

Cross-Site Compatible

Shared conventions and formats enable data comparison and synthesis across all four observatory regions.