It is based on the data created in the previous whitepaper "Usable Customer Intelligence from Social Media Data: Network Analytics meets Text Mining". Here text mining and network analytics techniques were combined to extend the feature set describing each forum users.

This feature set has then been the basis for this second analysis. Here we clustered the data using the k-Means algorithm and we identified a few clusters for neutral and inactive user, one cluster for very active and enthusiastic users (your superfans), a few smaller clusters for still active and still quite enthusiastic fans, and finally some clusters with negative fans and various degrees of activities.

A few interesting conclusions emerged from this study, like how to proceed with users from the different clusters or the concept of leadership and follower with respect to the concept of general activity. 

The whitepaper was developed with the KNIME Team and, like for the previous one, the pdf file and the KNIME workflows are downloadable from the KNIME whitepaper site.

A new KNIME whitepaper is available under "Resources"/"white papers" on the KNIME web site (http://www.knime.org/white-papers).
The title is: Usable Customer Intelligence from Social Media Data: Network Analytics meets Text Mining.

The paper shows  on one side how to process text data to extract keywords and how to use those keywords for a sentiment analysis; on the other side shows how to interpret the inter-user connections inside a forum to quantify the authority of each user. Finally, it shows how to put together these measures (authority measure and sentiment measure) to better position each user of the forum.

The whole analysis has been implemented with KNIME nodes, beside one R Snippet code to calculate the authority weight of each user. This paper then shows that KNIME is now able to handle network analysis and text mining inside the same platform.