Structured data organizes information into tables of rows and columns. For visualization work, tabular data is the everyday workhorse: column names identify dimensions and measures, while each row contains a data point. section 3.5 section 3.5.1
The crucial question is the table’s grain, or what one row represents. A row might represent one transaction, one day of summarized activity, or one account snapshot. Different grains require different treatment, and duplicated rows can send analysis sideways in a hurry. section 3.5.1
Working effectively with structured data means:
- Think in tables before building a visualization. section 3.5.1
- Identify the grain before interpreting, joining, or aggregating rows. section 3.5.1
- Use SQL to practice reasoning about rows, columns, joins, and aggregations, especially when working with BI tools. section 3.5.1
A fruit table
A fruit-tasting table uses one row per fruit observation, with columns for fruit ID, fruit type, taste score, and date eaten. Summarizing those observations produces a different table: one row per fruit type with its taste score. section 3.5.1
