Anyone have a good framework for sizing the residential solar market?

For state-level policy variability, I'd start with NREL's ReEDS model data and layer in net metering status + ITC eligibility by state to bucket states into tiers (favorable, neutral, restrictive). From there, multiply eligible housing units (RECS survey is great for this) by adoption rate assumptions per tier. Lawrence Berkeley Lab's 'Tracking the Sun' dataset is also super useful for calibrating your adoption curves against actual install data.