Methodology: Where Our Energy Cost Data Comes From
- Electricity rates
- EIA, per state
- Vehicle efficiency
- EPA fueleconomy.gov
- Incentives
- DSIRE + state/utility records
- Tax rules
- IRS source documents
What this page is for
Most energy-cost pages on the web do not tell you where their numbers came from. This one does. If a figure on SolarCostData looks wrong to you, this page should let you check it rather than take our word for it.
The four datasets behind every number
Residential electricity rates — U.S. Energy Information Administration (EIA). Rates drive almost everything else on this site: solar payback, the cost of charging an EV, what a heat pump does to a winter bill. We use the EIA’s per-state residential figures.
Vehicle efficiency and range — EPA. The EV and gas comparisons use the EPA’s fueleconomy.gov dataset rather than manufacturer marketing figures.
Incentives and net metering — DSIRE, plus state and utility program documents. DSIRE is the starting point; where a utility publishes its own tariff or program terms, we read those too, because DSIRE records can lag a utility’s own paperwork.
Federal tax rules — IRS source documents. We read the source rather than other websites. That is deliberate: most of the stale 2026 information in circulation comes from sites that copied each other rather than checking the primary document.
What our numbers are, and what they are not
They are estimates from public data with the formula shown. They are not quotes, they are not tax advice, and they are not a substitute for a site visit — the terms of use spell out that scope. Where a calculator lets you enter your own bill and rate, that version of the answer will always be closer to your reality than any national average we could print.
We also try not to round a modelled number into something that looks more precise than the data supports. If a figure is a range in the source, we keep it a range.
How often this gets refreshed
Rate and incentive data are refreshed when the underlying sources publish new figures, and pages carry an “updated” date so you can see how current the page you are reading is. Where a federal rule changes, we date the change on the page rather than silently overwriting the old position — you can see what applied before and what applies now.
Forward-looking fuel prices carry their own vintage: the winter-heating page and the two heating calculators’ optional preset use the EIA Short-Term Energy Outlook, September 2026 edition (released September 9, 2026; monthly national residential forecasts, averaged October–March). The STEO updates monthly; the official Winter Fuels Outlook supplement lands October 6, 2026.
State winter heating prices (the 20 cold-state pages and six fuel-versus-fuel pages under /winter-heating/) are last-winter measurements, not forecasts: electricity is EPM Table 5.6.A averaged over the October 2025–March 2026 editions; natural gas is the EIA Natural Gas Monthly residential price by state over the same months (1 Mcf = 10.38 therms — EIA’s residential price is average revenue per Mcf, so the high-volume winter months are the closest published proxy for a marginal rate); heating oil and propane are the EIA State Heating Oil and Propane Program weekly survey, averaged over its October–March season, with the final March reading shown alongside. Where SHOPP has no state series the PADD regional average is used and marked; where no regional series exists the national STEO figure is used and marked. The “this winter” column multiplies each state price by the national STEO change for that fuel and is labelled an estimate. Heat-pump rows use the NEEP cold-climate median COP for the 3-ton ducted class (3.20 / 2.17 / 1.90 at 47 / 17 / 5 °F). Heating degree-days are the STEO Census-division series. The build script (gen-winter-data.py) re-downloads all four EIA files, so the October refresh replaces every figure from primary sources.
The interactive tools, formula by formula
Every tool states its sources on its own page; this is the one-screen index.
- TOU rate simulator —
TOU = share×peak + (1−share)×off-peak; break-even share= (flat − off) ÷ (peak − off). Presets: 14 utilities’ own tariff sheets (schedule, window, effective date, link shown). - Bill jump calculator — identity
ΔBill = kWh₂×(r₂−r₁) + r₁×(kWh₂−kWh₁)with rates from your own bills; weather leg scales this site’s RECS/DOE-benchmarked equipment baselines by EIA STEO regional degree-day changes; state fallback uses EPM 5.6.A year-over-year. - Heat-pump sizing estimator — Efficiency Maine’s published
≥20 BTU/hr·sqftscreening rule and its 80–100 %-of-load band, checked against NEEP cold-climate class medians at 47/17/5 °F. A bracket; Manual J is the answer. - Heat-pump crossover — furnace
$/MMBtu = 10×$/therm ÷ AFUE; heat pump= 293.07×$/kWh ÷ COP(T)with COP interpolated between a named NEEP unit’s 47/17/5 °F points; monthly mean temp= 65 °F − HDD/day(STEO regional HDD). - Heat-pump performance check — rated draw
= capacity ÷ COP ÷ 3,412kW; the month’s continuous-run ceiling is the physics line your added kWh is tested against. - Rate tracker — 24 monthly EPM Table 5.6.A archive editions per state (first-published figures; revised where this site holds a newer print), US average alongside.
- Find your rate by ZIP — Census ZCTA→county crosswalk → EIA-861 service territory → you pick; price
= residential revenue ÷ residential sales(EIA-861, vintage labeled), never a tariff. - Energy map city layer — the 66 city pages’ own modeled payback and displayed rate; nothing recomputed.
Corrections
If you find a figure that does not match its source, tell us and we will check it against the source rather than against another website. Use the contact page . Corrections that change a published number get dated on the page they affect.
Every factual correction to a figure on this site is dated and described in the corrections log ; the accountability policy behind it is on the about page .
Production modeling (state and city pages)
State production bands are modeled with the PVGIS v5.2 PVcalc service (European Commission JRC) using the PVGIS-NSRDB radiation database — the same NREL satellite dataset PVWatts draws on for the Americas, so the irradiance input is identical and only the PV model differs. Assumptions: 4 kWp, standard c-Si, 14% system loss, roof mounting, 20-degree tilt, due south, horizon shading on; each state is the mean of a five-point cluster around its largest metro. Values are raw — PVGIS roof-mount figures run 1-3% conservative against PVWatts, and no correction factor has been invented to close that gap. Cross-checked against PVGIS-ERA5, which agreed within 8% in 42 states; ERA5 reads systematically high in the mountain West, so NSRDB is treated as primary. A separate real-roof factor (85% of the ideal) covers shading and orientation. PVGIS is free to use with JRC attribution requested; the underlying NSRDB is public domain (NREL/DOE). Municipal utilities and rural electric cooperatives are exempt from state export rules in most states and set their own terms.
Frequently asked questions
Are these figures quotes?
No. They are modelled estimates meant for sanity-checking a quote you have already been given, or for understanding the shape of a decision before you ask for one. An installer’s price depends on your roof, your utility and your local labour market, and no public dataset can know those.
Is any of this tax advice?
No. We read federal tax-credit status from IRS source documents and describe what those documents say. Whether a credit applies to your return depends on facts we do not have. Speak to a tax professional before you file.
How do you handle a rule that changed mid-year?
We date the change and say what applies before and after it. The largest example on this site is the 2026 status of the federal residential credits, which we read from IRS sources rather than repeating what other sites still say.
Sources: U.S. Energy Information Administration (EIA) electricity rates · DSIRE incentive records · public IRS/OBBBA guidance. Figures are modeled estimates, not quotes or tax advice. See our methodology.