What is Data Blending in Tableau?

We can combine information from two separate sources in a single Tableau worksheet or display by using the data blending tool.Unlike joining, which works on the row level and often duplicates data that repeats itself in several rows.

A primary data source and a secondary data source are used in data blending. As a result of this, more relevant data from the secondary data source is shown alongside primary data. We can create graphs and charts in a single sheet by combining data from both sources.

Data Blending in Tableau

The fundamental distinction between combining two data tables and blending two data tables is the phase at which aggregation of data happens. Like in joining, the tables are first merged and then the data is aggregated, resulting in some duplicating values.

Whereas, in blending, the tables are kept isolated at the database. Once the data has been compiled, it is sent to Tableau, where it is transformed into a single table free of duplicates. It can commonly happen that the secondary table has more than one equivalent value in its rows to the primary table.

When to Blend your Data in Tableau?

Data mixing tool in Tableau is very beneficial in the following instances;

 1. For example, you can't utilise cross-database joins in Oracle Essbase or Google Analytics because they don't support them (an extract only connection).

 Tableau allows you to import or connect to a variety of data sources and then use data blending to combine them.This enables you to combine a combination of data from multiple data sources on a single Tableau worksheet.

 2. When your data values exist at different levels of detail or are of various granularity, you should consider using data blending.

3. When working with enormous datasets, it's best to use data mixing. It is preferable to blend the data rather than utilise a join, as joins are cumbersome when dealing with large databases.

On the contrary, when we blend data, it collects the data first and then mixes it as required. It saves a lot of processing power in case of large data sets.

Compute Using - for table computations Tableau

Table calculations are specific sorts of calculated fields that compute values in a visualisation. In the current depiction, they are calculated based on what is now displayed and do not take into account measures or dimensions that are not visible.

You can use table calculation for this.

●      Transforming values into rankings.

●      Transforming values to show running totals.

●      Transforming values to show percent of total.

●      Transforming values to show % difference.

●      Transforming values to show moving average etc.

Addressing and partitioning

Table calculations need the use of all levels of detail, whether for partitioning (defining the data set on which the calculation is conducted) or for addressing (determining the direction of the calculation). Let’s get practical by designing a view.

Sample-superstore is the data source in this scenario.

●      Drag the Order Date to the columns.

●      Drag the Region and Segment to the rows.

●      Add Sales to the label shelf.

Next add a table calculation ‘Difference’ as shown below.

Table (across)

For each row, the Year (Order Date) difference is computed (by default), and this is done for each row (Segment).

 

It is possible to change the direction of our table computations using the 'Compute Using' option.

 

The following screenshots demonstrate how the various 'Compute Using' choices are implemented.

Table (down)

The calculation is computed down the rows Segment for every column YEAR(Order Date) (Order Date).

Table (across then down)

The formula is done across the columns YEAR(Order Date), then down a row Segment, then across the columns again for the complete table.

Table (down then across)

This calculation is computed down the rows segment, then across a column YEAR(Order Date), and then down rows again.

Pane (down)

This calculation is done down the rows Segment for a single pane.

Pane (across then down)

This calculation is done across the columns YEAR(Order Date) for the length of pane, then down a row Segment and then across columns for the length of the pane.

Pane (down then across)

Row Segment is used for the entire pane, then YEAR(Order Date) is crossed, and finally the entire pane is traversed one more.

Cell

This calculation is performed entirely inside the confines of a single cell.

Compute using specific dimensions.

This computation is computed within the dimension you provide.

 In the following view, the computation has been computed using dimension ‘Region’

Relative to

Table calculation options include 'Relative to', which provides the point of reference for computing your table calculation. Some of the alternatives you can use include, previous, next, first, last.

Tableau's Data Blending Restrictions

1. There is no way to publish a blended data source on the server as a single data source.

 2. Make your cube the primary data source.(In the case of combining data from a cube source with data from other sources.)

 3. It is necessary to constantly aggregate data in computations from the secondary data sources.

 4. You might have challenges while utilising non-additive aggregates like MEDIAN, COUNTD, etc while blending data.

Summary

This finishes our course on data mixing in Tableau. Here, we learned about the basic nature of data blending and how it works.

 However, if you’re looking for a reputable data science course in bangalore that covers all the fundamentals I suggest you take a look at a data science course in bangalore.Learnbay offers industry-recognized data science training. Machine learning, TensorFlow,Tableau, and IBM Watson are just a few of the topics covered in the courses offered by this institute.