The profusion of media data is growing at such a rapid rate that we face a future in which it threatens to become utterly unmanageable. In such circumstances, the answer lies in the ingenuity with which we apply technology to gain an understanding of exactly what is being said on a global scale.

 

But even when information from the media is filtered using search strings and logical operations, the final result can still require additional processing and refining.

 

Speaking at the AMEC Virtual Global Summit on Measurement, Nesin Ahmed, Service Delivery Manager, A Data Pro, delivers a presentation that explains how companies can better understand developing topics in the media using technology solutions.

 

Drawing on the example of the proliferation of pandemic topics in the media over the past year, he reveals that of the 400 million news stories across the globe, 80 million were on the topic of COVID-19.

 

Nesin emphasises the exact combination of technologies that his company uses to help to bring a 360-degree overview of the information that exists online and how audiences are reacting to that information.

 

“We use Machine Learning to tackle the huge amount of data, in a model known as Latent Dirichlet Allocation,” explains Nesin. “This is a predictive and statistical model which when run upon a data set consisting of documents analyses the prominence of the keywords. It creates topic clusters, then each document in the dataset is allocated into one or more topic clusters.”

 

The model that he outlines allows A Data Pro to track trending topics and pinpoint emergent topics as they start to gain traction in real time.

 

To learn more, this Machine Learning approach is combined with more traditional information such as structured data, for instance information about the news outlets and geographical locations. This allows A Data Pro to focus in on the topics that certain media are writing about, or gain insights into the specific conversations that are going on in specific places around the world.

 

Nesin concludes by commenting on the broader applications of such insights: “Having the ability to pinpoint emerging topics and trends in near real time can really enhance our analysis and can help with reputation management and brand management.”