text mining

In a recent article How Predictive Analytics Can Boost Your Social Media Campaigns, we saw that social media data is becoming a great source of data for understanding product trends by applying predictive analytics. We also emphasized that in order to stay ahead of competition, businesses need to make better use of the social media chatter to understand customer sentiments and customer’s needs.

Text mining is mostly associated with the analysis and insight generation from such type of unstructured data which is sourced from various social media like Facebook, Twitter, blogs, review websites, online newspapers, ebooks, flickr etc. While text mining for predictive applications such as customer churn require application of data mining algorithms, there are multiple ways you can perform basic text analysis without the need for sophisticated text mining algorithms. These methods are valuable for descriptive analysis of text data and for visual insights. Below we present two very common and useful tools for visual analysis of text data.

text mining

Word Tree: One of the most popular and useful method in this is the “Word Tree”. A word tree is a visual tool for unstructured text analysis from a book, article, speech or poem. It lets you pick a word or phrase and shows you all the different contexts in which the selected word appears. The contexts are arranged in the tree like branching structure to reveal concurrent themes and phrases. With the right phrase a word tree can sometimes reveal the heart of a dataset.

Word Cloud: Word clouds have also emerged as a straightforward and visually appealing visualization method for text. They are used in various contexts as a means to provide an overview by distilling text down to those words that appear with highest frequency. Typically, this is done in a static way as pure text summarization. Therefore there is a larger potential to this simple yet powerful visualization paradigm in text analytics.

Some of the applications of word cloud and word tree are

  • Patent analysis
  • Opinion mining
  • Investigative analysis
  • filtering terms in web based information retrieval
  • reviewing vocabulary of novel
  • filtering speech
  • exploring network of characters in novel

Application of this word cloud or word tree can give us a several useful insights about our unstructured data from social media or any available contents. Before we embark on more complex text mining, running such simple visual insight generating tools would be very effective.

Photo by Sixteen Miles Out on Unsplash

Originally published on Mon, Mar 17, 2014 @ 09:20 AM

No responses yet

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.