I'll systematically check for data quality issues across the key columns in this dataset. {"action": "tool_call", "reasoning": "Start by examining the measures and key dimensions for outliers and impossible values. I'll check Average Cost for two (has a suspicious max of 800,000), Aggregate rating (should be 0-5 range), Votes, Price range, and look for any negative values or zeros that might be problematic.", "tool": "query_sql", "source": "raw", "purpose": "Check for outliers and impossible values in numeric measures", "sql": "SELECT COUNT(*) as total_rows, COUNT(CASE WHEN \"Average Cost for two\" = 0 THEN 1 END) as zero_cost, COUNT(CASE WHEN \"Average Cost for two\" > 100000 THEN 1 END) as extreme_cost, MAX(\"Average Cost for two\") as max_cost, COUNT(CASE WHEN \"Aggregate rating\" < 0 OR \"Aggregate rating\" > 5 THEN 1 END) as invalid_rating, COUNT(CASE WHEN \"Price range\" NOT IN (1,2,3,4) THEN 1 END) as invalid_price_range, COUNT(CASE WHEN \"Votes\" < 0 THEN 1 END) as negative_votes, COUNT(CASE WHEN \"Latitude\" < -90 OR \"Latitude\" > 90 THEN 1 END) as invalid_lat, COUNT(CASE WHEN \"Longitude\" < -180 OR \"Longitude\" > 180 THEN 1 END) as invalid_lon FROM raw.\"Zomato Restaurant Dataset\""}