Kaveri Scraper

Sale-transaction coverage per seeded society — our scraper vs Propstack vs CRE, 2025 & 2026. Only genuine sales (consideration ≥ ₹1 lakh) are counted; distinct deeds by document number. Under each year: Counts (Propstack, CRE, Ours), then for each reference source — Ours-only (we have it, they don't) and Missing (they have it, we don't), matched by document number — against Propstack, CRE, and their union P∪C. % columns = share of the society's search tokens already searched, per search type: its own numbers (Survey, Khata), its per-flat ids (PID, RDPR) and the aux types over its numbers as one column; a number being re-checked counts as searched. Default order: most-scraped first; the picker filters to chosen societies. Ours counts the deeds the nightly linker put in the society (transactions.ks_project_id) — the panel below says how that link was decided.