17
SEP
2020

Tableau Extensibility Framework

Kommentare : 3

The first thing you should notice about this blog post is, that it has got a highly precise name “Tableau Wafer Extension”, but the word “Framework” attached to it. The reason is, this Tableau Extension solves a highly specific problem, but it can be leveraged to solve way more challenges!!! Therefore don’t look at it as a single highly specific Extension, but focus on the different modules and what they’re doing and think about your own scenarios where:

  • Using selected marks to trigger something outside of Tableau,
  • Generating images on-demand,
  • Visualizing images in a certain context,
  • Sending JSON data from your Extension to a WebService,
  • Calling a Python script from your Extension or
  • Writing enriched data back data into a database

 would be of value!

 

How can I get it?

 

Architecture

High Level Architecture

 

Demo

Demo

First 7minutes: Use Case Description + Wafer Extension Demo. The rest of the time is focusing on a Technical Deep Dive into it in order to let your tech-guys adjust it to your needs.

 

System Requirements:

    • Tableau Desktop (2020.2 and higher)
    • Port 5000 needs to be open/free (Flask Webserver)
    • Postgres database v11 (v12 not yet supported as of post date)
      • Postgis installed
    • Python 3.8 mit Anaconda 
    • Admin rights

 

Installation

Step-by-Step Installation

  1. Setting up the database with Ludwig Ehlert’s Postgres 11 install-setup walkthrough
  2. Download this Github repostory 
    • with the Tableau Extension
    • the TREX manifest file which points to the Tableau Extension
    • The v2020.2 Tableau Workbook
    • Python file with Flask Webserver
  3. Extract somewhere on your machine where you’d like to run the WebServer from in the next step (Ready. Set. Go!)
  4. Install Python 3.8 with Anaconda (leverage this ENV file for all the library dependencies: wafer_ext_env)

 

Ready. Set. Go!

Ready. Set. Go!

You can finally start your Flask WebServer which is part of the “json_io.py” script. For Windows open up Anaconda command line console as Administrator and type in the following commands (for Linux users: you know what you’re doing! 😉


cd C:\<THE PATH WHERE YOU EXTRACTED THE WAFER EXTENSION TO>

conda activate flask

set FLASK_ENV=development

python json_io.py
Now you should be all set! So please open up the Tableau Workbook within the 7Zip file and repeat what you can find within the demo video section.


Troubleshooting

 

Troubleshooting

    • Open up the “json_io.py” within a text editor and adjust the rows underneath “#####DB info” so that they do exactly match with your database setup.
    • If you run into issues please try to change the locale settings of the laptop you’re running the Tableau Workbook on to English (United States)
    • Make sure port 5000 is free
    • Check the logs within the Anaconda console
      • If you’re encountering a “ValueError: could not convert string to float:” within Anaconda’s debugging console simply change your workbook locale in Tableau Desktop to “English (United States)”
    • Debug the Tableau Extension similar to how I demonstrated it within this video:
    • With these steps in mind, please ask questions within the commenting area underneath if you’re having other issues in getting the Extension up and running

 

Timo Tautenhahn
Über den Autor
Timo originally coded his own BI-application by making use of different JAVA libraries. He Worked for IBM as BI-specialist and BA-architect before he joined Tableau. Favorite subject: JavaScript API
  1. Avatar
    Thomas Paschkowski Reply

    Awesome demo Timo!

    • Timo Tautenhahn
      Timo Tautenhahn Reply

      Thank you very much Thomas!

  2. Pingback: tableaufans.com » Tableau Extensibility Framework – More Use Cases

Kommentare

*

captcha *