Jitter discrete x values in SQL for Databricks AI/BI scatter

Databricks AI/BI scatter widgets have no jitter option, so points with the same discrete x (such as a 0-6 count) stack on top of each other. Spread them in the dataset SQL instead: n + ((ROW_NUMBER() OVER (PARTITION BY n ORDER BY y) % 4) - 1.5) * 0.11 AS x_spread, then format the x axis with decimalPlaces exact 0 so the ticks still read as integers. When shifting a generated dashboard layout with regex, check the resulting positions afterwards. One widget matched two patterns and was moved twice, which left an 8-row gap.