Monitoring workbook performance is an important part of creating dashboards. Larger data sources, busy servers and lots of worksheet actions can all contribute to slowing down the dashboard rendering and loading process.
The performance of your dashboard/workbook becomes increasingly important depending on the audience. Here’s a rough guide to different audiences and their patience for loading times:
CEOs/CFOs/the public: a few seconds
People interested in the topic (e.g specialists): 20-30 seconds
Data analysts: A few minutes
Some things that affect workbook performance are relatively out of our control, such as:
- Hardware
- Location of machines
- Network speeds
So, what can we control?
The workbook optimizer was released with Tableau Update 2022.1. From Server - Run Optimizer, you can quickly check things that might be slowing your workbook down and update them.
Here is a rough guide the things that could be slowing down your workbook performance...
Data Sources
Start with the data. Do you need millions of rows if you’re building something at the monthly level? Can you aggregate the data in Prep before loading it into Tableau? Generally, it’s better to have two data sources that are smaller and aggregated to the appropriate level than one larger data source. Aggregate to the level you need at the view if you have no filters on your dashboard.
Think about how many people are using the database at the same time for a live data source. Tableau is never ‘real-time’ analytics. You shouldn’t need calculations to be updating every 5 minutes. Do you really need a live connection over and extract?
Calculations
Table calculations are quicker than LODs. They only take into account what is in the view at that time. LODs query the data source then render. This could be the difference between 3 marks on your pane and 30,000 marks in the data source! Calculations that reference another calculation affect performance, although they are generally better for handover purposes. Consider recommending combining these calculations in one calculation after handing over your work.
Boolean calculations are quicker than strings.
Instead of: IF [Sub-Category]=’Copies’ THEN ‘True’ ELSE ‘False’ END
Use: [Sub-Category]=’Copies’
Filters
Too many filters can cause a slow dashboard. Try using multiple dashboards at different levels of aggregation to avoid having to use so many filters.
A rank of filters by their processing speed:
- Action filter - charts and interactivity (e.g clicking on a bar chart to filter a table)
- Wild-card filter - enter text to box filter values
- Apply-button - within the menu of the filter, have an option to ‘apply’ the filter. This could, however, save time in that the user is less likely to filter when they don’t need to.
- Only Relevant Values
Other quick tips:
- Hiding unused worksheets
- Removing unused data sources
- Clear downloads folder out.
- Have a column for every measure (long and thin, not short and wide!)
- Row for every date, not column.
- Use numerical parameters over strings
- Remove any unnecessary items from your view.