Text Analytics

Series 1: A re-emerging data trend

Background Information

With some experts in the analytics industry forecasting revenues hitting $18.4 billion by 2024, it begs the question what is text analytics?

Data mining can use traditional statistical and machine learning techniques to investigate and assess overlooked information that is hidden in the electronic text.

In fact, over 70% of an organizations data is unstructured, leaving text analytics a valuable tool in tapping into this goldmine of non-predefined data. Truth be told, text analytics is not a new topic. Text Analytics has been recognized and defined as part of business intelligence since 1958. Text Analytics is currently resurfacing for the following reasons: the ability to store lots of data, and the processing speed to do the analytics.

Text Analytics is applied to various industries including, business intelligence, research, exploratory data analysis, and investigation. Data mining tools extract this valuable text information, generate predictive models that provide a company user patterns and behavior, and then apply this information to an interested firm or organization.

With text analytics, an organization can take linguistic, statistical, and machine learning techniques to provide meaningful information that is hidden in all electronic text. This approach of analyzing data automatically assesses, analyzes, and acts upon the predictive models based on the information found in the electronic text. Uses of text analytics include call center logs, survey data, social media e-mail, loan applications, and insurance claims.

Text analytics encompasses three areas:

  • Information Retrieval – collecting textual materials via the web, database, file system or content management system for analysis.

Content Analysis consisting of:

  • Named Entity Recognition – statistical techniques that identify people, organizations, place names, stock tickers symbols, etc.
  • Sentiment Analysis – taking subjective (opposing factual) material and extracting opinions, moods and emotions. This information can form useful direction and approach in gauging sentiment to topic levels and concepts.

 

Implications

Text analytics – like BI dashboards, reporting, OLAP, and pivot tables – is one of the many business intelligence components, which extends BI from data in structured databases to the world of text data sources.

With astute analysis, text analytics can drive better business results through sound evidence-based predictive modeling. This is groundbreaking information retrieval that is made possible by sharp predictive accurate data mining tools. Methods of tapping into structured data work well, but with text analytics as a plethora of valuable information, it can greatly supplement numerical data for an overall acute analysis for superb business decision-making.