Companies are looking increasingly to take advantage of Big Data, especially textual information, those generated via user tools by web or desktop applications. The analysts specialized in this subject believe that 70% of information of interest to business are nestled in word documents, excel, email, etc. These data are not predefined in a model and cannot be perfectly stored in relational tables. They occur most often in the very free form, but contain dates, numbers, key words, facts that can be exploited.
A new challenge for companies in data analysis is to significantly advance the operation of this type of unstructured data. In terms of customer knowledge for example, it is possible to better use of the archives of business proposals and contracts, or listen to web conversations or take advantage of dialogues via email. Master relationships, including discussions about the company with its community of customers and stakeholders in its ecosystem, are very important for marketing today which has to change, push in that way by customers who largely use new technologies (mobile, social media).
The amount of this type of digital information usable is constantly growing, and as "manual extraction" of information is extremely difficult or even impossible on a large scale, so the use of specific computer tools for data processing unstructured text is required. Thus recently appears text mining tools, which automate the processing of large volumes of textual information, to identify statistically different topics raised and extract key information.
Text analytics techniques apply to the documents linguistic processing, including morphological, syntactic, semantic, and various other techniques of data analysis, statistical classification, etc. A major objective is to synthesize texts (classify, organize, summarize) by analyzing relationships, structures and rules of association among textual units (words, groups, phrases, documents). In the end it automates the production and the management of documents (including abstracts) and information (extraction, research, dissemination).
Text analytics has many applications for example in the field of customer relations, it allows in particular: exploring the contents of documents (e.g. open-ended questions in a survey, comments and complaints from customers, analysis of warranty claims) ; assign documents to predefined topics (redirect, mail filtering, organizing documents into categories, ranking contacts for call centers) ; compose text summaries (abstraction and condensation) ; examines texts by concepts, keywords, subjects, phrases to get results sorted by relevance like Google ; and finally increase the performance of predictive models by combining text and structured data.
To conclude, text analytics are a collection of technologies that detect the elements, or building blocks, within language, turning them into a type of data that can be manipulated and computed. To go further you can click on the link below to discover why it is necessary to use advanced analytical tools such as specialized Teradata Aster, to fully exploit unstructured data. Retail or Internet companies like Barnes & Noble or LinkedIn are already using these text analytics solutions in order to get a competitive advantage.