Commenting on various kinds of content on the Web (notes, posts etc.) is an integral part of social media, which emphasizes an interactive nature of online media and enables any participant to express opinion.
In our recent article we’ve already touched upon a topic of monitoring comments in mass media in the internet and their importance. They can sometimes provide much more food for thought and insights than an original article.
Tracking discussions in a variety of media is critical to a complete monitoring and analysis of the information field of the brand. This is complicated however by the fact that comments are reflected in search engines and a majority of monitoring platforms in the form of a general tape of mentions, i.e. in the form of a general incoherent flow of messages where it’s easy to get lost. But in terms of convenience of using the system and tracking people’s reaction to this or that information, it’s much more convenient to see the entire discussion connected with the original article/note.
In SemanticForce we’ve realized a special architecture to store, search for and visualize comments which can be shown below the original article or note they primarily referred to. Thus, you can see an overall picture of discussions developing before you from the title of the original post to the reader’s last remark.
The above format is used to represent comments for:
How does this algorithm work? If an article containing an object of monitoring is commented, you will be notified thereof by a special link below the text “show (hide) a comment”, with all left comments to be gathered there, whether they include a key word of monitoring or not. Indeed, when leaving a comment, you will not necessarily mention a brand or a product name, but can use only general words or pronouns. Thus, visualized comments help gather more mentions of the product than a conventional search does.
It’s necessary to note that visualized comments are especially useful for video – the entire chain of discussions below the video under tracking which sometimes evolves into a heated argument will be seen right in the system.
It’s also convenient to work with visualization of comments in social networks: comments to posts in VKontakte, for instance, are sometimes scattered on different pages, but a visualization algorithm allows gathering all of them in one integral thread of conversation.
We continue working to improve the mechanism, and a capability to define sentiment and subjects (tags) for each comment gathered in this way will be available shortly.