The data intelligence company analysed the sentiment of social media posts for one of South Africa’s largest beverage companies. It was able to do so with high accuracy at speeds that are unprecedented in the South African media intelligence industry.

This feat is made possible with NLP. Newsclip uses NLP to improve how its processing systems understand the language in media clips, including social media posts. It also enables automatic sentiment analysis, which allows companies like its beverage industry client to receive immediate insights into their media coverage.

Newsclip uses NLP to improve how its processing systems understand the language in media clips, including social media posts.
According to Catherine Dabbs, a representative of Newsclip, the company’s purpose-built NLP engine is being applied to process the media coverage of clients that generate large amounts of data.

“NLP breaks down large bodies of text to a granular level. It relies on the mechanics of language to understand the context of each word within every sentence and article.” The technology analyses text, like media clips, accurately because it is able to recognise when the names of people, places, and events are being referred to by synonyms or pronouns. The engine is also able to recognise slang and learn new forms of words as it processes, allowing it to evolve as the use of language changes.

“The result is near 100% accurate sentiment analysis on any form of media clip or media coverage, regardless of the amount of clips,” says Dabbs.

The value of accurate sentiment analysis

Accurate sentiment analysis of media coverage becomes vital when the analysis is being used to inform decision making.

Cindy Huber, marketing coordinator at Newsclip, explains: “Our clients often use our sentiment analysis to track public perception of their brand and adapt their communication strategies accordingly. If the analysis is not fully accurate, the company might miss vital customer feedback, which may damage their brand.”

She explains that sentiment analysis is used to track public reactions and attitudes towards new campaigns or product launches. If some media clips are analysed inaccurately, companies might not be made aware of small but significant problems with their products or campaign messages.

Sentiment analysis is used to track public reactions and attitudes towards new campaigns or product launches.
“Our near 100% accurate sentiment analysis keeps our clients updated on every aspect of the public’s sentiment towards their brands, regardless of the number of clips generated. The processing power we’ve put behind our NLP engine allows us to process thousands of media clips per minute, so you’re always in the know,” notes Huber.

In September 2017, the company’s engine processed 6.5 million media clips, with the potential to analyse much larger quantities when required to do so.

Addressing the need to make sense of digital

The digital revolution, which has now morphed into the digital age, has created a marketplace where customers demand accuracy and speed. Increasing data volumes, coupled with the major contribution social media makes to this escalation, means that tools which can sift through posts, status updates, and stories to find what is relevant to clients will become increasingly important to organisations that are always in the public eye.

Newsclip’s ability to conduct accurate sentiment analysis, at scale, is one way organisations are empowered to make sense of their place in the digital space.

For more information on Newsclip, visit www.newsclip.co.za or get in touch with the company on Facebook, Twitter or LinkedIn.

AI, machine learning, and NLP have made possible what was only imaginable a few years ago: extracting value from an ocean of information. Read how this technology works in our article, How AI makes sense of the media’s Big Data.