Written for the reader
Plain-language guidance for patients, clinical decision support for doctors, a jurisdiction view for agencies, from the same county forecast.
In development
Turn environmental data into the next step for a patient, a clinician and a health department.
Health Atlas turns environmental forecast into a public risk tier and the cited actions it calls for, written for a patient, a clinician, or a health department. The local climate record and projections sit alongside. Extreme heat is the first driver.
The decision What should a patient, a clinician and a health department each do on this forecast?
Primary audience Local health departments coordinating heat preparedness: public-health preparedness directors.
- Patients & caregivers
- Clinicians
- Local health departments
- State public health
Environmental advisories reach everyone in the same words. A patient, a physician and a health department each need a different next step.
The forecast is public. The action is not.
Good environmental health guidance already exists from the CDC, the National Weather Service, state health departments and state emergency management agencies. It is spread across advisories, fact sheets and clinical pages, and none of it is keyed to what the forecast says for one place this week.
- Guidance is written once, for everyone. A caregiver, a clinician reviewing a medication list and a county health officer read the same advisory and have to work out their own step.
- The forecast and the guidance live apart. Deciding whether a hot day is a Caution day or a Danger day, and what that changes, is left to the reader.
- Today’s conditions have no context. Whether this week is unusual for this place, and where it is heading over the coming decades, is not in the advisory.
One environmental forecast, three roles, every action sourced.
Keyed to a public scale
Each action starts at a named tier of a public scale (for heat, the NWS heat-index categories), so a Danger day brings different steps than a Caution day.
Sources attached
Every action carries its references: the CDC, NWS, state health and emergency management agencies, and peer-reviewed work. Nothing is the model’s own clinical advice.
From environmental data to cited action.
The same pipeline as every Tempra product: data, metric, tier, role-specific action, evidence. The climate record runs alongside as context, never mixed with the forecast.
County forecast
Hourly, county by county.
Heat index
Temperature and humidity combined into the heat index the NWS uses for advisories.
NWS heat tier
Caution, Extreme Caution, Danger or Extreme Danger, using the NWS categories.
Role actions
For each tier: self-care steps for patients, clinical steps for doctors, plan and outreach steps for agencies.
Evidence
Each action lists its sources: the CDC, NWS, a state health or emergency management agency, or a peer-reviewed paper.
Climate outlook
The local climate record and long-term projections, shown as context.
Where each role fits
Patients & caregivers
Plain-language guidance for the tier where you live, and a chat that explains it without diagnosing or changing medications.
Clinicians
The county forecast, its NWS tier and the cited clinician actions embedded in the assistant’s context for a patient visit.
Local health departments
Counties in the jurisdiction ranked by forecast tier, with the agency actions each tier calls for.
State public health
The same view rolled up from county to health region and state, for planning and outreach.
Health Atlas is decision support, not a medical device. It does not replace clinical judgment or official NWS and state health department alerts. The assistant is a language model run by Tempra Labs, and its output is labeled as model reasoning, not a statistical result. Any health-outcome analysis will be pre-registered and reported with its limitations. Demo data are fictional and labeled as synthetic.
Scoping a pilot.
One decision, one territory, a few weeks. The first conversation settles three things.
You bring
Approved guidance, your readers and your counties.
We provide
Local forecast, role-specific summaries and cited guidance for review.
We measure
Time to find applicable guidance, source traceability, usefulness to professionals.
In development. Official warnings and clinical judgment remain separate. Discuss a pilot