Independent product case study / Public-data intelligence
Voltline
Voltline combines the latest available EIA-930 submissions for Texas with monthly residential benchmarks for every U.S. state and D.C. A ZIP view maps 2020 Census ZCTAs to counties, uses 2024 ACS occupied-housing estimates for an explicitly modeled area-use range, and joins 2024 final EIA-861 utility context without claiming measured ZIP usage, a confirmed provider, or a current rate.
01 / Brief
EIA-930 exposes useful hourly balancing-authority submissions, while retail-sales data supports monthly state aggregates and EIA-861 supports annual utility context. None provides measured ZIP-level use or a single local rate, so any ZIP result must preserve the geography, vintage, calculation, and uncertainty behind it.
The product supports someone checking Texas hourly movement, inspecting a state benchmark, or comparing one or two ZIP-linked contexts with visible assumptions—not an operator, address-level utility locator, tariff engine, or household-bill estimator.
02 / Architecture
Each stage has a bounded responsibility. The interface preserves source context and leaves the final judgment with the person using it.
The server asks EIA for a four-day ERCO window containing demand, day-ahead forecast, net generation, and total interchange.
EIA API v2 primary · bounded four-day window · legacy fallback · 15-second timeoutRows are checked for a usable data array and finite numeric values, then sorted using their reported timestamps.
Invalid or absent demand ends the request with a visible errorDemand sets the current reporting hour. Forecast is matched to that exact timestamp; generation and interchange use their latest numeric observations and keep their own timestamps visible.
Latest demand + matching forecast + latest available adjacent signalsThe interface checks, in order, proximity to the displayed-window peak, forecast variance, and hour-over-hour movement before returning a neutral recent-range state.
95% of recent peak · 4% forecast variance · 5% hourly movementThe server treats the entered code as a 2020 Census ZIP Code Tabulation Area, retains every county overlap, and uses the largest land share as the primary county for a bounded service-territory lookup.
ZCTA proxy · county overlap shares · no USPS delivery or address-level claim2024 ACS occupied housing scales state residential use into a modeled ZIP-area range, while final 2024 EIA-861 records surface potential county-level utilities and their annual averages. A second ZIP adds a side-by-side comparison against an aligned state benchmark.
ACS estimate + margin of error · latest state month · annual utility contextEvery view keeps the source, vintage, reported period, geographic basis, limitations, and direct verification paths beside the result.
Hourly, monthly, annual, and five-year-estimate vintages · explicit non-tariff boundaries03 / Product decisions
Tradeoffs made visible
The context labels are fixed thresholds over retrieved values. They are easy to inspect and challenge, but they do not forecast conditions or infer grid reliability.
Forecast is compared only when it shares the latest demand timestamp. Generation and interchange may lag, so their own latest timestamps remain visible instead of implying alignment.
The route allows a ten-minute shared cache with stale-while-revalidate support. That lowers repeated upstream load while reinforcing that this is an hourly public view, not operational telemetry.
The ZIP view offers a useful estimate without relabeling it as observation: ACS occupied housing and a state per-customer benchmark produce an area-use range, while county EIA-861 matches produce potential utilities. Neither establishes a household’s usage, a confirmed utility, a tariff, a live supplier offer, or a ZIP-level rate.
Failure is a designed state
The route attempts the bounded EIA API v2 feed first, then a legacy EIA dashboard fallback. If neither is usable, it returns a no-store error; the interface clears stale results, announces the failure, offers retry, and links to EIA.
Voltline stops rather than filling the dashboard with inferred values. Forecast, generation, and interchange can remain explicitly unreported without invalidating a valid demand observation.
The client prevents concurrent loads and ignores responses from superseded requests, keeping late network responses from replacing newer interface state.
Voltline distinguishes a valid-looking code that lacks a represented 2020 ZCTA from a malformed input. When a required state benchmark is unavailable, it returns a clear unavailable state and never fills a missing benchmark, utility average, or ACS estimate with an invented value.
04 / Evaluation
This is the evaluation surface documented by the implementation—not a claim that every target has already been met.
Check that the latest reported hour and every adjacent-signal timestamp remain visible, and that caching never masquerades as real-time telemetry.
Exercise malformed arrays, nonnumeric values, missing demand, and individually missing forecast, generation, or interchange observations.
Verify sorting, exact forecast alignment, independent latest timestamps, recent-history ordering, and boundary behavior across reporting dates.
Test every threshold and its priority order, then confirm the copy never upgrades a pattern into an alert, forecast, or reliability claim.
Verify leading-zero and multi-county ZCTAs, land-share ordering, 2024 ACS estimate and margin-of-error math, EIA-861 joins, bundled-versus-delivery labels, and refusal to infer household usage, a unique utility, tariffs, or offers.
Verify that two ZIPs compare the same residential sector and aligned EIA month, that same-state results do not invent a ZIP winner, and that utility averages remain annual supporting context rather than current quotes.
05 / Provenance
The route retrieves EIA-930 region data for ERCO and exposes the official balancing-authority dashboard as the primary verification path.
The state explorer uses monthly residential sales, customer counts, and average retail price as state aggregates; recent months can be preliminary and revised.
Voltline uses Census statistical ZIP representations and county overlap, not an address lookup or a claim that every USPS ZIP is represented.
Occupied housing units and their margin of error support the modeled ZIP-area usage range; ACS housing counts are estimates, not electricity-meter observations.
County service-territory records identify potential utilities, while annual residential revenue, sales, and customer totals support bundled, delivery-only, or energy-only averages with their reported or imputed status preserved.
EIA explicitly states that it does not publish electricity consumption or average retail price by ZIP Code. Voltline therefore labels its area-use output as a model and its utility matches as county-level candidates.
Numeric validation, chronological ordering, signal selection, ZCTA lookup, ACS range calculation, utility-component separation, optional comparison, and cache headers are implemented locally in the site’s read-only API route.
Near-peak, forecast-variance, and hourly-movement labels are deterministic presentation logic—not EIA conclusions and not a model-generated forecast.
Implementation note
This case study was checked against the current Voltline client, versioned ZIP reference layer, and /api/grid-live implementation on this date. Live values, upstream availability, ACS estimates, and EIA revisions can change after the review; every result displays its own period and evidence boundary when run.
See the system under real conditions