If you're a health researcher thinking about incorporating climate or environmental data into your work, you're not alone — and the learning curve may be less steep than you think.
This post is a practical orientation for researchers who are new to climate data. It's not a comprehensive methods guide (those are coming in our Resources section), but it should give you a useful mental map for getting started.
Why Climate Data Matters for Health Research
Climate variables — temperature, humidity, precipitation, wind, air quality — are increasingly recognized as important determinants of health outcomes. Heat exposure drives cardiovascular and occupational health risks. Precipitation patterns shape vector-borne disease dynamics. Air quality affects respiratory health.
Yet many health researchers who want to study these relationships face a practical barrier: climate data is structured differently from health data, uses unfamiliar units and conventions, and comes from sources that health researchers may never have encountered.
Where Climate Data Comes From
There are three major families of climate data:
- Weather station data — ground-based instruments measuring temperature, precipitation, wind, humidity, and pressure at specific locations. High temporal resolution (often hourly), but spatially sparse. Quality varies.
- Satellite remote sensing — spaceborne instruments measuring surface temperature, vegetation, precipitation estimates, and atmospheric composition. Global coverage, but with spatial and temporal resolution trade-offs.
- Reanalysis products — hybrid datasets that combine station observations, satellite data, and atmospheric models to produce globally gridded, temporally continuous fields. Examples include ERA5 and MERRA-2.
Each has strengths and limitations. The right choice depends on your research question, spatial scale, and temporal requirements.
Common Challenges
The most common challenges health researchers face when working with climate data:
- Unit inconsistencies — Temperature may be in Celsius, Fahrenheit, or Kelvin. Pressure in Pa, hPa, mmHg, or inHg. Wind speed in m/s, mph, or knots. Our
weathertoolsR package handles these conversions automatically. - Temporal alignment — Climate data may be hourly or sub-daily, while health outcome data is often weekly or monthly. Aggregation decisions matter.
- Spatial alignment — Climate data may be point-based (stations) or gridded (reanalysis), while health data is typically facility- or district-based. Linking the two requires careful spatial matching.
- Derived variables — Many climate-health studies need derived metrics like Heat Index, Wet-bulb temperature, or Dew Point that aren't in raw data. Again,
weathertoolscomputes these from base variables.
How CGHI Can Help
This is exactly the gap CGHI's analytic infrastructure is designed to fill. Rather than every research team independently solving these data engineering challenges, we provide:
- weathertools — an R package for cleaning, converting, and deriving climate variables from raw data. Designed to be fast, reproducible, and hard to mess up.
- dataplane — an R package for creating self-describing Parquet files with embedded metadata, so datasets remain documented as they move between projects and teams.
- Data inventory — a growing catalog of climate, environmental, and health data sources relevant to our observatory regions.
- Methods guidance — tutorials and vignettes (in development) for common climate-health data workflows.
Next Steps
If you're ready to start working with climate data, here's a practical path:
- Browse our Data Inventory to see what's available for your region of interest
- Install
weathertoolsand try the getting-started vignette - Read the Resources section for currently available tutorials and orientation materials
- Use Contact CGHI for specific data or methods questions so your message can be routed to the right person
The barrier to entry is lower than you think. We're here to help.