DSS - > Tableau -> Tableau Server

emate Dataiku DSS Core Designer, Neuron 2020, Registered, Neuron 2021, Neuron 2022 Posts: 91 ✭✭✭✭✭✭

Hi All,

I am looking for a guidance /tips on how to set up in most efficient way a DSS output with Tableau, to give an end-user a best experience in using a dashboard.

So to sum up I have a DSS flow ready, and my output is not pretty "big" (~50k rows / 10 columns), on the other hand I have few calculations (LODs, Custom date filters) that might affect performance, this is why I am lookin for a best (in terms of performance) way to establish a connection between my data / tableau / server.

My data will be refreshed every month, so my first guess would be to use some kind of extract and then connect it to a tableau? Would that below is the best way to go? Also could someone explain me how this should work? I mean should I extract this data once a month to my Tableau server and then use this extracted data on the server as a datasource and connect it to a tableau dashboard?


Thanks for any tips.

Best Answer

  • AshleyW
    AshleyW Dataiker, Alpha Tester, Dataiku DSS Core Designer, Registered, Product Ideas Manager Posts: 161 Dataiker
    Answer ✓

    Hi @emate

    You could download the data to your computer and then upload it to Tableau Server as a data source or you could use the Tableau plugin to do that for you.

    Once the data is in Tableau Server however, what’ll make all the difference in terms of performance is how you’ve designed and built the workbook itself.

    I’d suggest ensuring that the calculations are optimized for speed, you’ve pushed calculations ‘lower’ (maybe even out of Tableau, if needed), choose at which level your filters should be applied (since they occur in different orders—LODs are great fun!), reducing unnecessary marks in a viz, etc to reduce query load.

    The Tableau Community site has a mountain of resources to help you optimize your workbooks and dashboards; I’d suggest heading over there too and looking at discussions around filters/calculations/performance!



  • emate
    emate Dataiku DSS Core Designer, Neuron 2020, Registered, Neuron 2021, Neuron 2022 Posts: 91 ✭✭✭✭✭✭

    Hi @AshleyW

    Sure, I will do that.

    As for this case, I already reduced dataset size and complexity to it's limit.

    I will look into the calculations and visualizations itself but I since, in this dashboard, the idea is to give end-user a lot of 'freedom' in terms of filters/custom dates are calculated dynamically, so I thought maybe I should also take care of the optimal way to set up this connection DSS -> Tableau, but now if I think about it, since the optimal way is to get extract, and I will be doing it once a month , DSS won't be some kind of limitation in terms of performance.

    Thanks for your answer!

Setup Info
      Help me…