\begin{tabular}{lp{10.5cm}rr} \toprule Rule & Description & Removed & Remaining \\ \midrule R0 & House listings (property\_type in house types) evaluated by Vrai Prix & 0 & 39,278 \\ R1 & Vrai Prix classifies the matched roll unit as single-family (unifamilial) & 6,041 & 33,237 \\ R2 & Listing geocode within 15 m of the roll unit & 12,636 & 20,601 \\ R3 & Roll unit is a single dwelling & 1,550 & 19,051 \\ R4 & Complete roll fields: year built, floor area, land and building values & 404 & 18,647 \\ R5 & Listing year built (if any) within ±5 years of the roll year built (match sanity) & 331 & 18,316 \\ R6 & At least 4 image URLs & 598 & 17,718 \\ R7 & Asking price within [100,000, 3,000,000] & 209 & 17,509 \\ R8 & Sources with no structured building fields and few photos excluded: ['kijiji', 'fb\_marketplace'] & 3,633 & 13,876 \\ R10 & Image host permits automated retrieval (robots.txt): hosts with 'Disallow: /' excluded & 5,287 & 8,589 \\ R9 & One listing per roll unit (duplicates across sources: keep the listing with most images) & 844 & 7,745 \\ \bottomrule \end{tabular}