Tableau Data Analyst cheat sheet
Tableau Data Analyst · Keep the key ideas within reach.
Data Preparation and Connection
Live connection runs every query against source; extract (.hyper) copies data locally and can refresh on a schedule or manually.
Join types: inner keeps matching rows only; left keeps all left rows; right keeps all right rows; full outer keeps all rows from both sides.
Relationships (noodle) are logical, lazy, and keep tables at their own level of detail; joins are physical and merge tables into one flat result.
Blending requires a primary source plus at least one secondary; linking fields must match on name or be manually mapped, and only the primary's dimensions can filter the view by default.
UNION appends rows (same column structure required); JOIN adds columns. Pivot converts wide crosstab columns into rows (Columns to Rows) or rows into columns (Rows to Columns).
Data Interpreter is a checkbox on the Data Source page that strips titles, footnotes, and merged headers from Excel/PDF-style crosstabs before Tableau reads the table.
Changing a field's data type in the Data Source grid affects calculations and defaults; geographic roles (Country, State, ZIP) enable mapping.
Extract filters can be added at extract time (Data > Extract > Filters) to reduce .hyper size; hidden fields are not removed from the extract unless filtered out.
Default aggregation is SUM for measures; dimensions are typically disaggregated, and 'Aggregate Measures' off in Analysis menu disables measure aggregation.
Exploratory Analysis and Data Visualization
Marks card controls Type (bar, line, square, circle, text, pie, etc.), Color, Size, Label, Detail, Tooltip, and Shape.
Show Me appears when one or more fields are selected and recommends a chart type; it is a shortcut, not a requirement, for building views.
Continuous fields produce axes (green pills); discrete fields produce headers (blue pills). Converting a date to discrete creates a header per period.
Calculated field syntax: [Field], functions like SUM(), AVG(), COUNTD(), IF/THEN/ELSEIF/END, and LODs {FIXED [dim] : SUM([measure])}.
Table calcs (RUNNING_SUM, WINDOW_AVG) are computed after aggregation and depend on addressing/partitioning via Compute Using.
Quick Table Calculation options include Running Total, Percent of Total, Difference, Percent Difference, Moving Average, and Rank.
Reference lines (Analytics pane) mark constants, averages, or medians on an axis; box plots and trend lines are also added from the Analytics pane.
Filter order of operations: extract filters, data source filters, context filters, dimension filters, measure filters, table calc filters. Context filters are evaluated before other dimension filters.
COUNTD counts distinct non-null values; COUNT counts rows including duplicates; ATTR returns a value only when all rows in the group share it, else '*'.
Dashboard and Story Design
Dashboard size options: fixed (exact pixel dimensions), range (min/max), and automatic (fills available space). Choose fixed for predictable layouts.
Layout containers (horizontal/vertical) control how sheets and objects resize; tiled items fill space, floating items are positioned freely and can overlap.
Actions: Filter, Highlight, Go to Sheet, URL, and Parameter actions. Filter actions can run on hover, select, or menu, and can target specific sheets.
Dashboard device previews (phone, tablet) let you design separate layouts; the phone layout is optional and does not affect the desktop layout.
Stories are sequences of story points; each point captures a sheet or dashboard snapshot. A blank point can be added and populated later.
Dashboard filter cards use 'Apply to Worksheets' to control which sheets receive the filter; 'All Using This Data Source' applies broadly.
Parameters are workbook-level and not tied to a data source; they can drive calculated fields, filters, reference lines, and bins.
Tooltips can embed sheets (Viz in Tooltip) to show a miniature view on hover without adding it to the dashboard.
Sharing, Governance, and Analytics Best Practices
Publish to Tableau Server/Cloud; permissions are set per project, workbook, or data source, with roles like Viewer, Interactor, Editor, and Publisher.
Extract refresh schedules run on Server/Cloud; live connections query the source at view time and require source credentials or embedded credentials.
Row-level security can use user filters (User Filters) or calculated fields referencing USERNAME() or ISMEMBEROF() groups.
Data source filters are applied at the source level and propagate to all sheets using that source, unlike sheet-level filters.
Workbook version compatibility: newer Tableau versions can open older workbooks, but older versions may not open newer ones; use File > Save As with a target version when needed.
Publishing options include 'Include external files' (for images, local extracts) and 'Embed credentials' to allow refresh without prompting.
Subscriptions deliver snapshots on a schedule; they use the permissions of the subscriber, so a subscriber without access sees no data.
Performance: reduce marks (aggregate early), use extracts for large sources, avoid overly wide tables, and prefer context filters over many dimension filters when order matters.