The other ledger for Georgia: more high-hazard dams in poor condition than any other state on AP's national count, Atlanta congestion worsening two years running (65 hours, then 75), 22 rural hospitals modeled vulnerable, and 8,500 MW of forecast load growth the state's own regulators think is overstated — both numbers printed. And bridges at just 2% poor, reported exactly as measured. Every figure is sourced, dated, and labeled — counted, modeled, or assessed.
How to read this page: the class chip on every measure is the honesty label. Counted = an inspection or enumeration record. Modeled = a study's projection or risk score — printed with the modeler's name. Assessed = a structured survey of required investment — needs as documented, not measurements of decay. Figures keep the form their source gave them; years are printed on every measure.
Click any category to expand it — every measure with its figure, what it means in plain language, and its source.
Dams. The ≈200 poor-condition count is the Associated Press's 2019 analysis of the National Inventory of Dams — the freshest national per-state analysis retrieved; the 542 Category I total is ASCE Georgia's 2024 figure. Both carry their years wherever they appear.
Food. The tract-level food-desert share (20%) is from a 2014 USDA Atlas analysis — the last tract-level pass found; Feeding America's modeled food-insecurity rate (16.7%, 2024) is the current-need companion. No newer statewide Atlas figure was dressed up as current.
Power — a contested forecast, printed both ways. Georgia Power's 8,500 MW load-growth forecast (2025 IRP, via Utility Dive reporting, flagged in the data) is paired with the PSC staff's recommended 2,000 MW reduction from the same proceeding. The disagreement is the information.
Broadband. The 482,374 unserved-location count is the state's own 2021 address-level map — the pre-BEAD baseline against which the $1.307B allocation was sized; programme progress since is tracked on the Built side as projects are funded.
Traffic. INRIX hours are metro figures (Atlanta); no honest statewide congestion total is published for Georgia, and none is printed. Both published years (2024: 65 hours; 2025: 75 hours) are shown because the direction matters.
Classifications normalized (editorial, disclosed). Three researcher labels were harmonised so the same evidence wears the same chip in every state: the Public Policy Foundation shortage estimate and TRIP's pavement share are 'modeled' (as on the NC/TN/SC/VA pages), and the PSC staff forecast-reduction is 'modeled' rather than 'assessed'.