Hey Tableaunew learning, welcome to the retail community.
There are likely multiple factors in play here. Here's a list of questions you can evaluate.
1: How large is the data set? Is it a live connection or an extract? Is the level of granularity in the datasource much lower than what is being presented in the dashboard?
2: What type of data source are you connected to? (excel, csv, Teradata, Oracle, etc.)
3: What types of visualizations are being used?
4: Are you using context filters or just standard quick filters? (Tableau filter order of operations can be found here: Tableau's Order of Operations )
5: Have you taken a performance recording to isolate the long running queries/actions? If not, here is how you do that: Create a Performance Recording
6: How are your dashboard actions configured?
You likely are unable to share the .twbx for the file you have built since it is corporate data, but any additional detail you can provide about the data structure/dashboard layout would be very helpful.
Answers are below:
1) Size of data is .9 million. It's an extract, I am showing data for each Month.. however SKU data is present in filters.. so my understanding is granularity of data is SKU.
2) data source is SQL server
3) It's 2 tables and 2 graphs (Bar and Line graph)
4) there is only one filter and rest other filters i have converted into Sheets and using those sheets as filters
5) I haven't taken performance recording yet
6) there are 10 filter actions... one for each sheet (used as filter).
Let me know if that helps in providing any insight on performance
Thanks, that's helpful. One thing that catches my attention is you mentioned there is SKU data in the filters. That's a likely culprit if there is a SKU list being populated in a filter. That would definitely cause slow load and also slower response times, because each action you take would cause it to repopulate.
Is there any way you can share a sanitized screenshot(blurred logos/other sensitive or identifiable data points)? I can likely interpret what the intended purpose/action of each sheet is and that could help isolate a particular issue.
Performance recording will also be your best friend here. It'll call out whether the lag is due to load or query executions.
Thank you for providing screenshots. To me it looks like you're trying to satisfy about 10 use cases with this dashboard(which isn't necessarily bad, just requires a bit of strategy). The way I would describe this to my users is, instead of trying to be everything to everyone in one dashboard, we can strategically construct a guided-analysis sequence that allows us to tell both macro and micro stories.
Here's what I mean by that. Maybe your initial dashboard is at Department and Class level with a view on each product level and also one for months. This can then filter down to a similar dashboard at the next lowest levels as you guide your users through the analysis. An alternate thought would be to create drillable hierarchies for your product levels. Then using that with "keep" only, you could create a solid flow.
The reason it's slow the way it's constructed today, is that by filtering on everything/having every level visible in a sheet simultaneously, Tableau is having to build essentially all .9M rows you mentioned as individual marks every time you take an action.
Hope this helps!
So you mean for e.g. build initial level dashboard for each Department then as user choose the department ... take 'em to another dashboard with filtered view on class and so on for SKU..? based on selection in one dashboard it will refresh other dashboard?
Yes, i'd advise building an initial dashboard at a higher level. Maybe a filter on the highest level of your product dimension, with the 2nd/3rd highest levels being displayed in the views. Then use a filter dashboard action to drill down into additional dashboards that are more focused toward understanding the lower levels. (i.e. a bar chart works well if there there are 4-5 Departments, but won't be as effective for something like SKU where there are 1000's(scatter plot could be better here).
Can you explain me what is the mean by Sorting Data Event while generating Performance Recording in Tableau
Thanks in Advance
Saw your post over on the main forum. It's a good question that i don't have a quick answer to without a bit more context. Looks like the long sorting events are occurring on a single worksheet. Could you provide some detail on what's going on w/ that worksheet? I.e. is it a tall and wide grid? how many rows/marks are included? is there a table calc rank function in play? etc.
Thanks Larsen Rennie