Huge data flows can yield virtual gold mines of insights, predictions and other analysis, but what happens when the text-based data is in multiple languages? To address Big Data’s Big Language problem, customer engagement provider SDL has announced that its cloud-based BeGlobal machine translation (MT) solution can now integrate with text analytics processing, boosting the ability to track customer sentiment and business trends in multiple languages. 

A component of the company’s Language Platform, BeGlobal, enables a real-time translation of social media communications, web content and other text-based data. On a monthly basis, its MT solution translates billions of words from structured and unstructured data, involving over 80 language translation combinations.

Single Language

The general approach presented by BeGlobal is to translate the multilingual text into a single language, such as English, for processing. Translated, the data become grist for the mill of text analytics solutions, including predictive analysis, sentiment analysis, search or other uses. One of the applications is e-discovery, such as when lawyers need to unearth every relevant document, even if they’re in different languages.

BeGlobal provides an open API for integration into an existing system, and the translation is offered through a scalable, software-as-a-service model. HTTPS protocol and other features are provided for encryption and secure ID.

To indicate the size of Big Language, SDL notes that, every day around the world, there are 175 million new Tweets, 2 million blog posts, more than 68 million posts on Tumblr, 60,000 new websites, and, every 60 seconds, nearly 300,000 Facebook status updates. Only one-fifth of Internet users are native English speakers.

Multinational Virtual Assistants

SDL said that language support has been difficult for text analytics companies to provide, because of the R&D and hardware requirements. According to a 2012 report from Gartner, only about one-third of text analytics companies support multiple language translation, and most of those only several languages.

In its announcement, SDL cited several actual use cases of its technology. An unnamed sentiment analysis company, for instance, employed the company’s machine translation to determine different global regions’ views of clients’ products or brands. Similarly, other unnamed companies can now offer virtual assistants for multinational customers or text analytics of international news broadcasts.

Image courtesy of Lonely Walker (Shutterstock)