Problem
Raw month-to-month usage comparisons are misleading because weather varies year to year. With full historical weather (#1) we can normalize.
Proposal
- Compute heating degree-days (HDD) and cooling degree-days (CDD) per bill period from daily temps (base 65°F, configurable).
- Add a "usage per degree-day" metric and a weather-normalized usage view so months/years are comparable and real efficiency changes are visible.
- Keep the math pure + unit-tested (
lib/series.ts or lib/weather/).
Acceptance criteria
- HDD/CDD derived per month; a chart shows weather-normalized usage.
- Hand-calculated test for the degree-day rollup.
Blocked by #1 (needs daily historical temps).
Problem
Raw month-to-month usage comparisons are misleading because weather varies year to year. With full historical weather (#1) we can normalize.
Proposal
lib/series.tsorlib/weather/).Acceptance criteria
Blocked by #1 (needs daily historical temps).