Last fortnight we brought you some relatively minor updates, with features such as a new sentiment classifier for Portugese not quite warranting a blog post on the revisions.

Never fear however, as it simply meant we were working extra hard on bringing you two great new features this weekend instead.

This round of updates has a number of little improvements across the tool, but the most significant of the lot are the new Authors feature and a radically updated Topics component.

New Look Topics

We’ve scrapped our old topics system entirely, as we were aware of some of the shortcomings. Replacing the previous component is a box-fresh redesigned version, which works entirely on-the-fly, making it perform much faster.

Before, the topic tab may have taken up to a couple of minutes to populate, with the Brandwatch app working hard to sift through all them mentions to discover recurring themes.

Topics now operate on a sample of the data instead, allowing superfast processing of the topic cloud and real-time updating of terms with the greater accuracy than ever before.

This has the added benefit of improving upon the old system in another way, as topics were not automatically deleted previously, whereas now they are constantly updated.

Another novelty with this update is the inclusion of custom-created tags and categories in the topic cloud. This allows Brandwatch users to see all of their categorisation features in one place, and presents the topics within categories too.

Topics are scaled in size depending on volume of mentions within that topic, so more popular themes will appear larger. Automatically generated topics will appear in dark grey, tags are pink and categories are blue.

The terms can also be scaled according to burst, which is the rate the topic has emerged over time.

The colours can also be modified so that sentiment can be reflected in the presentation, or so that Brandwatch’s colour palette is adhered to.

Clicking on an individual topic will allow users to see all the mentions of that query that are about that topic.

Another new feature to the updated Topics component is the ability to see the topics of a topic, namely the terms that are commonly found within another term by ‘zooming in’. Here we’ve shown all the topics within ’22 computer’, which itself is a topic of the query for Raspberry Pi.

One final note with this new feature is that all existing topics are about to be replaced. If you already use Brandwatch you will have already heard from us about transferring any manually created topics over to the new system, but if you’re reading this on the March 2nd then make sure you are aware of this update going into the weekend.

Topics can no longer be manually created, in favour of our much more flexible Rules system. Get in touch if you’d like to know more about how to use this feature.

Authors

The other exciting new addition to Brandwatch rolling out over this weekend is an evolution of our top tweeters component. Leaving the ‘Top Sites’ tab behind, we now have a dedicated section for the authors of mentions online.

For this low-volume example query, we can see that Dr Roy Schestowitz is the most popular contributor to discussion of Raspberry Pi, behind admin, which is a common username in forums.

This section identifies the authors which appear the most frequently across the web for that query. It will also unite matching usernames across multiple platforms and sites as a single user. Here’s PaulWordman talking about Raspberry Pi in two different places, recognised by Brandwatch as a single author.

The authors can also be listed and ranked in a table format. Here users can see the individual mentions themselves, as well as the sentiment breakdown and other metadata.

The Authors tab also displays the top tweeters, the top forum users (which can also be shown for a solitary forum, which would be interesting) and top bloggers too.

We’ve also added an authors’ column to the mentions page, an author filter to the controls panel and even established a new Boolean operator, so that searches can be performed using ‘author:’ to isolate mentions of a single (or multiple) authors.

We’re always working hard on finding ways to improve Brandwatch. If you have any comments about these new features, or any ideas for improvements to our app, please do get in touch with us: we’d love to hear what you think!