This article provides a high level overview of using Tableau Prep, Tableau Desktop / Online and Tableau Bridge to setup a data pipeline
Introduction
My recent assignment in my Master degree was involved in developing a set of Pharmacy analytic dashboards for a non profit cancer treatment center. The goal of the project is to develop a set of dashboard and allows their teams to analyse and monitor the key areas (workforce management, patient experience, dispensing and financial management) in order to achieve better patient care.
Given that they have no pior knowledge on data science and so my team had decided to use Tableau software. The diagram below illustrates the end to end flow:
We used the following Tableau software:
- Tableau Prep for data merging, cleansing, feature engineering and generating the output files to be used as data sources in the dashboards.
- Tableau Desktop for developing the initial dashboards
- Tableau Online for viewing the dashboards online
- Tableau Bridge for updating the data sources
Here is the high level flow:
The input files were generated by their Pharmacy system (Merlin), either manually or automatically. We then created a scheduled job to kick off the Tableau Prep process for merging, cleaning and generating the output files.
Tableau Prep builder provides a feature where you can use their in built Data Interpreter to analyse the input source and incorporate with the dataset.
Tableau Bridge will then be running on the same machine, it has a scheduling function that allows user to periodically transfer the latest output files and update the data sources in Tableau Online. You can setup the scheduling details in Tableau Online once the bridge is connected to your account:
The Tableau Bridge can then be started from the Tableau Desktop.
The Tableau Online was used because the client want to show these dashboards on their TV. If there is any issue transferring the file, an email will be sending out to the user so that they can investigate the issue.
I found the Tableau components are very user friendly. They all have a GUI that allow non technical users to setup the data pipelines quite easily, without doing any coding. They also provide an easy way to extract the input data from different format and very easy to integrate all these components together. Having said that, Tableau Prep takes a quite a long time to load and Tableau Bridge supports only Windows. I would recommend if the dataset is not big and your client /team just want focus on their expertise, using Tableau to setup a data pipeline is actually a good option.
