Search this blog

Tuesday, 8 March 2011

Generating reports in a RapidMiner process

Here's an example of a RapidMiner process that generates reports. One report is a PDF, the other is a static Web page.

The basic flow is to start with a "Generate Report" operator. This sets up the name, type and location of the report; in this case a pdf file. From then on, the name of the report is used by subsequent operators to add things to it. In the example, the next operator creates a "series multiple" plot of the attributes of the iris data set plotted against the label. From there, a page break is added and finally another plot is added, this time of the cluster determined from a k-means clustering algorithm. The final report is located in c:\temp\Iris.pdf.

Multiple reports can be created so it would be possible to have many reports created in a single process.

In the example, a "generate portal" operator creates another report and similar graphs are reported to it. In this case, a static web site is created at c:\temp\RMPortal\irisweb.html. In the example, a tab is created that contains text and it is shamelessly gratifying to note that this can be raw html.

Presumably, the enterprise versions have richer graphics and more control. RapidAnalytics would allow reports to be created in accordance with a schedule thereby creating a standalone Web site acting as a reporting portal.


  1. Replies
    1. Hi,thanks for the example, it´s help me to understand the reporting management in Rapidminer. However, when I ran the example, the pdf was properly created, but none html file was created. The log window, shows the following message:

      Dec 1, 2013 1:14:22 PM WARNING: Problem during finishing the process: The reporting caused an error: Failed to create Table Of Contents.

      Could you help me with this issue?

      Thanks and best regards.

      Marcelo Saparrat

  2. Hello Marcelo,

    It's probably because the folder into which the HTML is written must exist first. On Windows it's c:\temp\RMPortal - you can change this of course.



  3. Thanks for this process example it been helpful, I want´t to know if there is a way to format the output of a column as percentage or money

  4. Hello Marco

    I would hazard a guess that it might be possible to use some sort of xslt stylesheet solution. However, this is "strong tobacco" if you know what I mean so I don't have an example I can share. Motivated by your question, I will have a think but I can't promise a result.