A new palette of social nodes has been added to the extensions hub.

For many use cases like customer satisfaction analysis or social media marketing response analysis it makes sense to utilize the social media application development interfaces. However, to do so in IBM SPSS Modeler’s visual interface you would need to go into programming with R or Python. To provide you a little more convenient access to social media data, we are now offering nodes with an easy to understand user interface in our extension gallery.

Facebook Extension
There are three new nodes to access public data from Facebook:

  • Facebook Posts: Import Facebook posts from a public page.
  • Facebook Comments: Import Facebook comments from a public post.
  • Facebook Reactions: Import Facebook reactions to a list of public posts.
  • Facebook Nodes
    Check out the github for all the details.

    Twitter Extension
    You have the following options to access Twitter data:

  • Twitter Trends: See what’s going on right now at a specific location or worldwide.
  • Twitter Search: Search twitter for a specific keyword.
  • Twitter Users: Get all details about users who tweeted about a keyword.
  • Twitter Timelines: Get all tweets of a list of users.
  • Twitter Nodes
    Check out the github for all the details.

    13 comments on"Social Nodes for SPSS Modeler"

    1. Hi, how could I dowload these nodes?

    2. There is a bug in the Twitter Trend node. It always shows trends for lat = 48 and long = 9. Actually the values inputed on the dialog are not used by it. It is not using the values of “latitude” and “longitude” variables defined in the syntax, it uses these two constants instead.

      • JonnyLangefeld October 07, 2016

        @George, I you are right, I fixed the bug in my repository. So via the github link above you’ll get the most recent version. I’ll make sure it gets updated in the extension hub, too.

    3. I have the question concerning the Facebook Posts. The node is able to download the post from my own Facebook page, or is able to download any posts from any Facebook user, groups, profile? F.e. If I put the keyword “IBM”, the FacebookPosts node downloads any post from any other Facebook profiles? I did not find it on your webpage.
      Thanks for your answer.

    4. JonnyLangefeld October 18, 2016

      @jozefK using the node you have the view as of ‘you’. Meaning anything you can see when you log into Facebook, you can also see with the API.
      The Facebook Posts node loads public posts from public pages. Because IBM is a public page, you can load all the posts from the page. You would do this for example to get the post IDs, with which you can load reactions of users. See the documentation here: https://github.com/jonnyenglish/SPSSFacebook/tree/master/Facebook%20Posts

    5. Ricardo Abreu December 09, 2016

      Hi,

      I have SPSS Modeller 18 for Mac in trial version, can i install this Nodes for Facebook API?

      Thank you

    6. Hi,
      I run the Facebook Posts get the following error
      “Data and data model does not match.”

      Would you mind to teach us how to solve it?

      • MateuszLapa1986 February 22, 2017

        Hello,

        I have the same issue with Facebook Posts … I am connecting to Pui Yu’s request.

        Thank you very much for your help.

      • Goh Kai Song May 07, 2017

        I encountered the same problem too.
        Seems like the RFacebook package’s getPage function now returns 11 fields instead of the original 10 fields.
        That’s why you get a mismatch between the Data and Data Model.

        Here’s how to resolve this.

        1. In Modeler, go to Extensions > Custom Node Builder Dialogue
        2. In the resulting window, select File > Open, then navigate to the directory where you have the “Import_Data_from_Facebook.mpe” file. Open it.
        3. You will be asked to select the .cfe file. Since we are fixing the Facebook Post node, select “FaceBookPost.cfe” and click Continue.
        4. On the top right panel, you will see the R script, 2 changes are required.
        a. In line 34, between var7 and var8, add a line to accommodate this new field, I called it var7a.

        var7a <- c(fieldName="Story", fieldLabel="", fieldStorage="string", fieldMeasure="", fieldFormat="", fieldRole="")

        b. In line 39 where the metadata for the data model is created, add this new field (var7a) between var7 and var8.

        modelerDataModel <- data.frame(var1,var2,var3,var4,var5,var6,var7,var7a,var8,var9,var10)

        5. Save the mpe file.
        6. Reinstall the extension.

        Cheers!

        • Hi there,

          Thanks for the fix. I applied it and it worked.

          Could you please provide a fix for the FacebookComments node as well?

          I also seem to be having the same issues with the Twitter nodes.

          Thanks

          • Goh Kai Song July 13, 2017

            Sorry for late reply. I don’t visit here often. Here goes.

            Follow steps 1 to 3 (except that you will select FacebookComments.cfe at step 3)

            Step 4: 2 changes are required in the R script.

            a. In Line 29, between var5 and var6, add var5a with the following line.

            var5a <- c(fieldName="CommentCount", fieldLabel="", fieldStorage="string", fieldMeasure="", fieldFormat="", fieldRole="")

            b. In line 31 where the metadata for the data model is created, add this new field (var5a) between var5 and var6.

            modelerDataModel <- data.frame(var1,var2,var3,var4,var5,var5a,var6)

    7. coachhelensmith January 28, 2018

      Are there any nodes for Instagram, Snapchat, LinkedIn, or Personal Facebook Profiles?

    Join The Discussion

    Your email address will not be published. Required fields are marked *