The needs below are documented, not rhetorical: a 195,462-home deficit for Michigan's poorest renters, at least $39.6B of water work priced by the University of Michigan, 42% of major roads poor or mediocre, ten high-hazard dams in poor condition six years after Edenville — and a $1.559B broadband build that has not yet connected its first BEAD location.
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.
Michigan is published at full coverage: all nine categories measured, 19 measures total. The state's stories are specific — a housing deficit concentrated in metro Detroit, a water backlog priced a decade after Flint, roads and bridges among the nation's worst by TRIP's count, and two peninsulas' worth of rural hospitals, food access and broadband distance. Where a Michigan-specific source exists at editorial standard it is used: NLIHC for Michigan, the University of Michigan's 2026 water assessment, TRIP's April 2026 Michigan report, EGLE's dam figures and NTIA's approved Michigan BEAD proposal.
Honesty labels do the work they do on every state page. A counted deficit (housing) is not a modelled rate (food insecurity) is not a needs assessment (water, dams) is not a utility's forecast (power). The hospital category prints two different things side by side and never blends them — a Senate-requested risk screen that flagged 3 hospitals, and 5 actual closures over 20 years. The power category is one utility's disclosed pipeline, labeled as such; Michigan's other big utility, Consumers Energy, publishes no directly comparable data-centre figure, so none is invented for symmetry.
Two deliberate omissions, stated plainly. First, no food-desert count: no verified statewide count of Michigan residents in USDA low-income/low-access tracts was available at this page's standard, so the food category measures food insecurity itself and says so in its definition note. Second, traffic is metro-only: TTI measures Detroit and Grand Rapids, and no honest statewide delay average exists, so none is printed. Flint's remaining ≈500 lead lines are carried as context inside the water measure rather than as a separate row, because the statewide figure (≈245,000 lines) is the sourced measure.