What are Data Connections, Data Sources and Data Models in Tableau?
In general when we are starting in our Tableau journey we often directly jump into the visualization part without pausing and trying to understand the importance of the data that we connect to in Tableau.
Unfortunately this is how I started my Tableau journey with an eagerness to build interesting views rather than understanding how the underlying data interacts with Tableau.
The purpose of this article is to provide a brief understanding of the three key terms when working with data in Tableau
1. Data Connections
2. Data Sources
3. Data Models
For each of these terms a visual summary will also be provided to better explain the concepts.
Data Connections:
The left pane of the data source page displays details about the data that Tableau Desktop is connected to. Connections is present on top of the left pane. Below this the related details of the source file are present e.g: file name, sheet names etc.
For file-based data, the left pane might display the file name and the worksheets in the file.
For relational-based/ databases server data, the left pane might display the server, the database or schema, and the tables in the database.
We can also use the left pane to add more connections to the data source to create cross-database joins (using the Add button).
The term Connection has another reference in Tableau.
In the top-right corner of the Canvas area in the Data Source Page , we can select the connection type as either Live or Extract.
Data Sources:
A Tableau data source may contain multiple data connections to different databases or files.
A Tableau Data Source is the link between the source data and Tableau.
A Tableau data source consists of the below mentioned components:
a) The actual data (either as a live connection or an extract) e.g. Embedded Flat Files
b) The connection string to the underlying database or the reference to a Tableau virtual connection
c) The names of tables or sheets containing data
d) Description of Fields
e) Any Metadata changes (the customizations) done on top of the data to work with it in Tableau. These customizations may include things like how the data is combined and metadata such as Calculations, Renamed fields, Data Type changes, Hierarchies, Groups & Folders and default formatting etc.
Hence there is a difference between the Tableau data source (Tableau’s access to the data and all the customizations that may be involved)
and the data source itself (such as a database or Excel spreadsheet).
A published data source contains all the details that anyone would need to start building visualizations/views in Tableau.
Data Models:
Every data source that we create in Tableau has a data model.
The data model is a diagram that tells Tableau how it should query data in the connected database tables.
A Tableau Data Model consists of the below mentioned components:
a) Connection information to the underlying data source
b) Queries that are to be executed to retrieve the data
c) Any Metadata i.e. the data about the data e.g: Renamed fields, Data type changes, Hierarchies, Groups & Folders, Custom Calculated Fields etc.
The tables that are added to the Canvas in the Data Source page will create the structure of the data model. A data model can be simple, i.e. a single table or it can be complex, with multiple tables that use a combination of data modeling options like relationships, joins, and unions.
The data model in Tableau (Version 2020.2 and later) consists of two layers
The default view that we first see in the Data Source page canvas is the logical layer of the data source. The data between the tables in the logical layer are combined using relationships (sometimes affectionately called “Noodles”).
The next or inner layer is the physical layer where we can combine data between tables using joins and unions. Each logical table may contain at least one physical table.
Hence the logical tables act as Containers for the physical tables.
The Data Model is used as the foundation for creating all further analysis in Tableau.
Once these key terms are understood it will help us in working better in Tableau and eventually aid in building better visualizations to address the business/end-user requirements.
I hope this article is useful for all Tableau users especially for anyone just starting their Tableau journey!!!
References:
“Part 1: Data Modeling on the Tableau Platform — Chapter 1 Introducing Data Modeling in Tableau” & “Chapter 2 Licensing Considerations and Types of Data Models” from the book “Data Modeling with Tableau” by Kirk Munroe
