A Tableau Prep Flow is a visual data preparation workflow built in Tableau Prep Builder that cleans, shapes, and combines data before it reaches Tableau Desktop or Tableau Server for visualization. This guide explains how Prep flows work, when they are the right tool, and how they fit into a governed analytics architecture.
A Tableau Prep Flow is a visual data preparation workflow created in Tableau Prep Builder that cleans, shapes, combines, and outputs data for use in Tableau Desktop, Tableau Server, or Tableau Cloud. Where Tableau Desktop and Server are for visualization and analysis, Tableau Prep is for the data preparation step that precedes visualization — transforming raw source data into analysis-ready output tables.
What Tableau Prep Does
Tableau Prep Builder is a canvas-based tool where data preparation steps are connected in a flow — a directed acyclic graph of data transformation steps from input sources to output destinations.
A Tableau Prep flow might:
- Connect to multiple data sources (Snowflake tables, CSV exports, Salesforce extracts)
- Clean a messy source (fix date formats, trim whitespace from string fields, replace null values with defaults)
- Reshape data (pivot columns to rows, aggregate to a higher granularity, union multiple similar tables)
- Join multiple tables (combine orders with customer attributes and product descriptions)
- Apply calculated fields (compute derived columns, apply business logic)
- Filter rows (remove test records, exclude irrelevant date ranges)
- Output to a Tableau .hyper extract or a database table
The visual canvas shows the lineage of each step — you can see at a glance that this Output node receives data from this Clean step, which in turn comes from this Join, which joins these two Input sources.
When Tableau Prep Is the Right Tool
Tableau Prep is designed for self-service data preparation — business analysts and BI developers who need to prepare data for Tableau without writing SQL or using a separate ETL tool. It is the right tool when:
**The data needs cleaning that cannot be done in Tableau Desktop's calculated fields:** Missing value imputation, string parsing (extracting a zip code from an address string), reformatting dates from ambiguous formats — these are preparation steps best handled before the data reaches the visualization layer.
**Multiple tables need to be joined for analysis that cannot join them efficiently in the data source:** If a Tableau live connection's multi-table join is too complex or slow, pre-joining in Prep and outputting to an extract is faster.
**Data needs to be reshaped:** Pivoting a wide dataset (one column per month) to a tall format (one row per month) before Tableau can analyze it correctly. Tableau Prep's Pivot step handles this visually.
**Incremental or scheduled preparation is needed:** Published Prep flows on Tableau Server can run on a schedule, refreshing the output extract automatically. This enables batch data preparation workflows managed through the Tableau platform.
Tableau Prep vs. dbt
Tableau Prep and dbt address overlapping problems — both transform data — but from different perspectives:
**Tableau Prep** is a self-service visual tool for analysts. It connects to data sources, transforms them visually, and outputs to Tableau formats (.hyper extracts, published data sources). No SQL knowledge required. The flow runs in Tableau's infrastructure. Best for: analyst-owned data preparation for Tableau-specific consumption.
**dbt** is a SQL-based transformation tool for data engineers and analytics engineers. It transforms data within the warehouse using SQL models, materializes results as warehouse tables, and tests and documents every model. Best for: governed, production-grade transformation logic producing warehouse tables consumed by any downstream tool.
For production analytics where multiple teams and tools consume the same data, dbt is the appropriate choice — the transformation logic is code-reviewed, version-controlled, and produces governed warehouse tables. For individual analyst or departmental data preparation work that feeds only Tableau, Tableau Prep is a lower-friction alternative.
Tableau Prep in a Governed Architecture
Tableau Prep flows published to Tableau Server run as scheduled jobs, output to published data sources. This enables Prep to play a role in a governed Tableau architecture — but with important caveats.
Prep flows that produce published data sources can be discovered and used by multiple Tableau users, which sounds like governance. But Prep flows are not code-reviewed, not version-controlled by default (without external source control integration), and their logic is not testable with automated tests the way dbt models are. For data that feeds executive reporting or financial close, the governance requirements of dbt are typically necessary.
For departmental analytics where the analyst owns both the data preparation and the dashboard, Prep flows are an efficient, self-contained solution.
Performance Considerations
Tableau Prep runs transformations in its own compute environment, not inside the database. This means large datasets take longer to process in Prep than they would as a database-executed SQL transformation. For source tables over a few million rows, Prep flows can be slow.
The practical guidance: use Tableau Prep for the pre-processing and output-to-extract workflow for medium-sized datasets (under ~10M rows). For larger datasets, perform the heavy transformation in the warehouse (via dbt or SQL views) and use Prep only for final light shaping before the extract.
Our Tableau consulting services covers Tableau Prep flow design, extract refresh architecture, and integration with dbt for governed data preparation. Contact us to discuss your Tableau data architecture.
A former Microsoft data architect audits your data foundation, identifies your top priorities, and sends you a written plan. Free. No pitch.
Book a Call →