4 Replies Latest reply on Jan 18, 2017 7:24 PM by Sarah Murphy

    Increasing processing speed of polygon maps

    Sarah Murphy

      Hi there,

       

      I am mapping all of the US Census block groups within a single state (Montana) as a chloropeth map. I downloaded the shape file and then followed the steps outlined here to create a single csv file where one row = one path. This results in a file that is over a million rows -- plus the ~35 columns worth of other data in the file.

       

      Unsurprisingly, the data takes a very long time to load any time that I try to filter (or do anything else, really). I have made sure that I have the "extract" option ticked off any time that I edit the data connection. Additionally, I have made sure that all calculations and merges between data sets were done outside of Tableau to help reduce the processing time.

       

      Is there something that I'm missing here? I know that there are many publicly available Tableau polygon mapping files, and these do not seem to have long processing times when I download them and click around.

       

      Apologies for not including a workbook here -- the data are somewhat sensitive. If it is too hard to understand my particular problem, I can try to manipulate the dataset to make it less sensitive.

        • 1. Re: Increasing processing speed of polygon maps
          Jamieson Christian

          Sarah,

           

          You said you created a CSV for the paths, plus the ~35 other columns of data. There are 2 thoughts that come to mind:

           

          1. Did you create an extract (TDE) from the CSV file? CSV is a slow data source to traverse, and you may find that Tableau can traverse its own TDE extract format much more efficiently. (EDIT: I just realized that you explicitly indicated that you turn OFF the "Extract" mode. You probably want to do just the opposite. Turn it on!)
          2. You may get a substantial performance boost by separating the "path" data and the "data" data and use data blending. Make the "path" data your primary data source, and the "data" data your secondary data source. That way, Tableau won't replicate the "data" data across every single row of your "path" data.
          • 2. Re: Increasing processing speed of polygon maps
            Chris McClellan

            There's 2 ways I can think of to attack this:

             

            1. If you're building the CSV yourself, can you reduce the number of points in your polygons ? that will make the drawing faster because there's less points to draw (I've seen that at a client site before where the drawing was a lot faster as they made the polygons "rougher" but didn't lose much detail because the maps were small in the dashboard)

             

            2. Have you seen Tableau 10.2 ? It's only in beta, but you can open a shape file directly and it's a lot faster than the old approach.  (http://beta.tableau.com/ ) remember this is BETA SOFTWARE, but a good view of what's coming very soon.

            • 3. Re: Increasing processing speed of polygon maps
              Sarah Murphy

              This is helpful!

               

              1. Did you create an extract (TDE) from the CSV file? CSV is a slow data source to traverse, and you may find that Tableau can traverse its own TDE extract format much more efficiently. (EDIT: I just realized that you explicitly indicated that you turn OFF the "Extract" mode. You probably want to do just the opposite. Turn it on!)

              Ha! This was poor word choice on my part -- by "ticked off," I meant that I had checked the box to turn the extract option ON. But what you're saying about the TDE is exactly what I'm getting at -- is there an additional step that I should be taking, or does turning on the extract option create the TDE file? I know that I've seen a dialog box pop up to save the data as a TDE when first loading a new data source, but not when simply editing the connection (i.e. when I need to connect to an updated version of the data).

               

              You may get a substantial performance boost by separating the "path" data and the "data" data and use data blending. Make the "path" data your primary data source, and the "data" data your secondary data source. That way, Tableau won't replicate the "data" data across every single row of your "path" data.

              I will try this. I had originally joined the two data sets myself outside of tableau because I read another old forum post that said this could help -- but I guess maybe it did just the opposite.

               

              Thanks again!

              • 4. Re: Increasing processing speed of polygon maps
                Sarah Murphy

                Thanks! Rougher polygons is something I may consider if all else fails.

                 

                I hadn't seen 10.2 yet, but I'm relieved to see that they are adding that feature. I was very surprised that Tableau didn't have a better way of dealing with shape files!