Role of Text Analytics In Social Media

In Social Media it is not only about reach, likes, retweets, and follows anymore. The spectrum of data is growing day by day and companies are facing difficulties in gaining the true value from social data. Though the current listening methodologies and strategies help organizations to understand customer pulse, there is a lot of valuable information left unnoticed by of listening experts. This happens because of of the amount of noise and real data harnessed around a brand or product. With the help of text analytics social media data can be sliced and diced to understand different aspect of consumer feedback. 


Gaining Structured Insight to Unstructured data

According to Wikipedia, The term text analytics describes “a set of linguisticstatistical, and machine learning techniques that model and structure the information content of textual sources for business intelligenceexploratory data analysisresearch, or investigation”. 

 Text analytics helps you quickly analyze thousands of social data to identify the customer issues and concerns that demand immediate attention. Text analytics has multiple applications in analyzing text but in this article we will be looking at the following aspects

  • Theme identification: Conversations can be auto-categorized and organized into relevant categories based on automatically detected themes and content.
  • Sentiment Analysis: Determine customer sentiment, on product or brand level.

Gaining Structured Insight to Unstructured data

Theme Identification: Text analytic techniques are useful for sorting through and categorizing large volumes of content. In Text analytics Words are treated similar to other forms of Data. In a text data set, Words are counted, grouped and summarized with statistics. As said before text analytic techniques are quite good at text categorization. 

Classification algorithm looks at features, such as the frequency of words in a document relative to the frequency of terms in all documents.  These techniques are also used to group similar documents even without predefined categories. Using this we can also identify emerging topics around a brand or product. 

Sentiment analysis: Text analytics also have some value in extracting the users emotions or opinion or attitude and classify accordingly but there are some limitations in this area. Sentiment analysis is one of the common text analysis task used by social media researchers in recent times. In common sentiment analysis is categorized as positive, negative or neutral.  Simple text analytic techniques, such as counting positive and negative words, can give reasonable results in many cases. Sentiment is classified based on the list positive words or negative words updated into the system. In such cases this techniques will not differentiate between sarcasm and genuine comments. 

Tools For Text Analytics

Text mining computer programs are available from many commercial and open source companies and sources.

Commercial Tools Open Source Tools
AeroText  Carrot2 
Angoss  GATE 
Attensity  OpenNLP
Autonomy  Natural Language Toolkit (NLTK) 
Basis Technology  RapidMiner
Clarabridge  UIMA
Endeca Technologies  Language R
Fair Isaac  The KNIME
IBM LanguageWare KH Coder
IBM SPSS Modeler Premium The PLOS
Inxight  Weka
Thomson Data Analyzer   

In the next article i will try to explain this technique using WEKA Data Mining tool. I will try to explain as comprehensive as possible. Please share your thoughts on this topic, i would love to learn it from you. !Thanks for reading !