Data Preparation and Connection
Connecting and shaping data
This domain tests whether you can get clean, correctly typed data into Tableau and know how Tableau treats it once connected. Understand the difference between live connections and extracts: extracts are snapshots that can be refreshed and are faster for large or remote sources, while live connections query the source each time. Know connection types (single table, custom SQL, and relationships vs. joins). Relationships (the modern model) keep tables logically separate and aggregate at the level of the sheet, reducing row-duplication; joins physically merge rows and can inflate totals when keys are not unique. Know when each is appropriate.
Data types, roles, and cleaning
Be fluent with data types (string, integer, float, boolean, date, datetime) and the discrete vs. continuous and dimension vs. measure distinctions. Know how to change a field's role, apply default properties, split fields, pivot, use the Data Interpreter, and handle nulls. Understand data source filters versus dimension filters versus table-calc filters - only source filters reduce the data pulled in.
Traps to watch
- A join that duplicates rows and silently doubles a SUM.
- Confusing a blended (secondary) source with a joined one; blending aggregates to the linking field.
- Assuming a filter placed on a worksheet reduces the underlying extract.
- Forgetting that changing a field to a dimension vs. measure changes aggregation.
Rule of thumb
Before trusting a number, ask: at what grain are the rows? If the join key is not unique on both sides, totals will be wrong. Validate by comparing a COUNTD of the key to the raw row count.