In my experience I couldn’t find a way to apply these two transformations natively in Tableau Prep (or that doesn’t require many additional steps) but I was able to integrate a custom Python script to meet my need. Nevertheless, two additional transformations are needed in order to leave the dataset ready to train models (transform the categorical variables into a numerical format of 1 and 0 to finally normalize them). At this point you can add an output process that would apply all changes to the dataset and export the result in CSV format or as an extract from Tableau Desktop to create reports. As can be seen, all steps could be performed in a minimum effort with a set of clicks and Tableau Prep has fulfilled its functionality (clean the data before creating reports with Tableau Desktop). Up to this point the dataset is completely clean and can be used for pattern analysis and reporting. Section 4: Adaptation of the dataset for modeling The formula used in the calculated field was: IF ISNULL() THEN ELSE ENDįinally, the column transformation process ends with the elimination of the field with the passenger ID and ticket number (Similar to the Cabin elimination step). To do this, I decided to replicate the cleaning process that I once did in Python to the popular Titanic dataset being careful to the point where the tool may fall short and if it is really compliant enough to apply to a larger project.Īgregation and Join steps in Tableau Prep As a consultant of this tool, I was then in the duty to explore its potential, to know its advantages and its real capacity in order to evaluate if it is viable to present it to the clients within their BI projects. Regardless of the recent changes, Tableau has made to its Tableau Desktop product in order to improve performance with large volumes of data, in most cases, it is necessary to add an ETL process such as Talend or Pentaho prior to analysis and reporting with Tableau Desktop.Ĭonscious of the previous, last year (2018) Tableau released to the public the product Tableau Prep Builder with the intention of providing a Drag & Drop tool prior to data exploration with Tableau Desktop. In my job as a BI consultant with Tableau, I’ve heard quite a lot of the phrase “Tableau is not an ETL” where I’ve had to agree most of the time. Tableau Prep, how powerful it is, where it fails and where its greatest advantages lie.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |