Introduction
Filters in Tableau allow users to refine data visualisations, focus on specific insights, and to create interactive dashboards. This post will explore the several types of filters in Tableau, focussing on their order of application and when to use each.
Types of filters
There are six main types of filters in Tableau, each playing a specific role in refining data and shaping visualisations. Below is an overview of these filters and how they fit into Tableau’s order of operations.
Image Source: https://help.tableau.com/current/pro/desktop/en-us/order_of_operations.htm
1. Extract filters
The main purpose of an extract filter is to reduce the size of the data before loading it into Tableau.
Example:
If you are working with a massive worldwide dataset, but you only need Australian data.
Drawbacks:
Once applied, you cannot access the excluded data without refreshing or recreating the extract.
2. Data Source filters
Data source filters are applied after the data has been loaded into memory. Their main use is to restrict data access for security or performance reasons.
Example
In an Australia-wide healthcare dataset, patient records could be restricted to only those within the current state for privacy regulations.
Drawbacks
Limits access for other analysts necessitating custom data source filters to be applied.
3. Context filters
Establishes a subset of data on which downstream filters depend, which can also improve performance.
Example
If you need to find the top 5 bedding products by sales, a context filter can be used to select the bedding products prior to applying the top 5 filter.
Drawbacks
May slow performance if applied to complex calculated fields or large datasets as they are applied every time the data or filter is updated.
4. Dimension filters (non-aggregated fields)
Filter categorical data. Filtering is applied at the row level before aggregation.
Example
Filtering a “Country” dimension to only display “Australia” results.
Drawbacks
Rows with null values in the dimension will be excluded, and applying dimension filters to high-cardinality fields can be resource-intensive.
5. Measure filters (aggregated fields)
Filter numerical data based on conditions or ranges.
Example
Filtering sales greater than $1000.
Drawbacks
Filtering on aggregated values can lead to unintended exclusions, as they are applied after the aggregation.
6. Table calculation filters (post aggregation)
Filter data at the visualisation level.
Example
Show top 5 performing products based on their rank in a table.
Drawbacks
Limited in scope, as they only affect the current visualisation, which can lead to inconsistencies when combining views or datasets.
Conclusion
Understanding the capabilities, potential drawbacks, and order of operations of Tableau’s filters allows for effective data management, insightful visualisations, and optimised dashboard performance.