About LoadLens
Adaptive load forecasting for
rural electric cooperatives.
LoadLens is a research platform that applies regime-aware adaptive ensemble machine learning to distribution-level grid load prediction. It continuously grades its own forecasts against real outcomes, re-weights its models based on recent performance, and publishes the full evaluation receipt — including failure modes — on a pre-registered methodology page.
The Problem
Rural electric cooperatives serve 42 million Americans across 56% of U.S. land area. For decades, their load forecasting tools worked well enough — daily and seasonal patterns were predictable, and static models could track them.
Four forces are making rural grid load fundamentally non-stationary: EV adoption, distributed solar generation, extreme weather volatility, and shifting rural demographics. Static forecasting tools fail at regime transitions — precisely the moments where accurate forecasts matter most.
EV Adoption
Neighborhood-level charging clusters double household load overnight.
Distributed Solar
Midday troughs and evening ramps that break traditional load curves.
Weather Volatility
More frequent extremes. Static 20-year-average models mispredict at the worst moments.
Demographic Shifts
Remote work, changing agricultural patterns, and data center buildout shift rural load profiles.
Market
832 rural electric distribution cooperatives in America, organized under NRECA. Most run on billing and CIS platforms from NISC or SEDC — their forecasting modules were designed for a stationary grid.
Enterprise analytics from Itron or Innowatts cost $200K–$2M/year and target investor-owned utilities. LoadLens fills the gap: adaptive forecasting at a cooperative-friendly price point, designed to complement the NISC and SEDC ecosystem.
Team
Champlin Enterprises is a premium software engineering consultancy founded in 1998. Twenty-eight years of production software engineering, including senior work for Wells Fargo and Fortune 500 consumer brands.
The adaptive ensemble framework that powers LoadLens is the same architecture behind Vantage AI, validated across two non-stationary domains — demonstrating cross-domain generalization.
Methodology & Integrity
Every accuracy claim on this site is backed by a single command, php artisan loadlens:eval, run against a versioned, pre-registered protocol. Holdouts, baselines, regime stratifications, and falsification conditions were declared in advance — not retrofitted to make a number look good.
The methodology page publishes the full receipt: the per-regime MAPE breakdown, calibration coverage on the probabilistic intervals, the regime-detection log, and an open list of known issues we're working on. It is updated automatically whenever the eval re-runs; nothing on it is hand-edited.
Research Context
LoadLens is a research platform for regime-aware adaptive ensemble forecasting applied to distribution-level grid load prediction.
All validation data on the Learning page is generated from real PJM Interconnection hourly demand data sourced through the EIA Open Data API. Eval receipts on the methodology page use the same source.