It is not uncommon for users to face performance issues while consuming dashboards deployed in Tableau Server. This may be experienced in the form of longer load times and/or longer interaction response times. In this blog we will be looking at a few ways to optimize performance of Tableau dashboards.
1. Set the dashboard layout to a fixed size
When creating Tableau dashboards, avoid choosing automatic size for dashboard rendering.This deteriorates performance as the dashboard has to render itself according to resolution of the end user device, which may vary.
Instead set the dashboard to a fixed size or any of the predefined screen sizes that is going to be compatible with the screen resolutions of most users. This ensures that there are no run-time resolution adjustments during consumption, and your dashboards will load faster.
2. Try to minimize mark points
Mark points refer to each object or value showed in a dashboard. A dashboard with high mark points (for E.g. a Tabular report) takes longer time to load, especially with a live connection.
You can enhance performance by utilizing charts to represent the same data. If you have a chart with many mark points, you may also restrict the charts by limiting the display to TOP/BOTTOM N dimension Values.
3. Minimize usage of Quick Filters
Avoid using too many quick filters in a dashboard. Let’s say you have 5 filters in your dashboard. Every time a filter is applied, Tableau sends a query to the data source to retrieve data points relevant to filter applied. Therefore, applying five different quick filters leads to Tableau sending query to data source five times which slows retrieval of data and display of Dashboards with filtered data.
Use Dashboard Actions instead of Quick Filters for better performance. This helps in faster rendering of Dashboard and enhances its look and feel. Use Quick filters only for the dimension or measures that are not used in the Dashboard charts
4. Minimize the number of charts used
Having many charts jumbled into a single dashboard directly leads to increased rendering and layout computing time. Especially when filters are being applied, the filtered data must reflect in all the charts present in Dashboard
The best practice is to use not more than four charts in a dashboard.
5. Retrieve only what is needed from the Data Source
In general, only fetch data that you need in the dashboard. For example, while using Extract connections, ensure that only relevant data is loaded into the Extract by utilizing appropriate Extract filters.Once dashboard is built in Tableau Desktop use the ‘hide unused columns’ option to remove columns from extract that are not used for building Dashboards – this reduces size of the extract and enhances performance.
6. Optimize your backend configuration (hardware & software)Ensure robust hardware, memory & storage for your backend/database to support faster data retrieval. This is especially more important for Live connections. The columns that are used frequently while building visualizations or as filters can be indexed in the database for better performance (Note: Indexed column should not be used as context filter)
Finally, avoid using Tableau Desktop and Server in the same machine as it may lead to performance issues unless there is enough hardware present to balance the load.
* * *
Learn more about Visual BI’s Tableau consulting & end user training programs here.