Brandwatch + Cision
By Giles PalmerFeb 26
Sometimes the greatest discoveries come from approaching problems with a different perspective.
Famously, Penicillin was only discovered after a petri dish was left out by accident.
The microwave’s culinary potential was only realized after someone noticed a chocolate bar melting after exposure.
And, Kellogg’s Corn Flakes only became a thing when Kellogg’s brother cooked some wheat that should have been thrown away.
By looking at problems from a different perspective, we’re often able to discover something new.
To help Brandwatch users discover new insights, we’ve created a new visualization.
The Brandwatch topic wheel lets users view the key topics and subtopics within conversations they’re interested in, segmented by a unique parameter (e.g. gender, sentiment, account types etc.).
The wheel is segmented into three layers. The inner layer shows the parameter you wish to break your data down by (e.g. gender).
The second layer will show the key topics associated with that parameter (e.g. males talk most about ‘speed’ and ‘acceleration’ in relation to cars).
The third layer shows the subtopics that relate to the key topics in the second layer (e.g. ‘Tesla Roadster’ and ‘Ford Mustang’ in relation to ‘speed’).
It provides a new view of your data that reveals richer insights.
Take a look at these charts. One is a simple word cloud for Tesla with positive and negative conversation highlighted in green and red.
It’s difficult to uncover an insight here.
What are consumers angry about? What are they happy about? How do they view Elon Musk after he appears on Joe Rogan’s show smoking marijuana?
It’s tough to tell.
Take a look at the same data visualized in a topic wheel.
Now we discover something interesting. An overwhelming volume of negative conversation seems to be about an Ex-Tesla employee being fired for failing a drugs test, while their CEO takes drugs live on a show.
Most other wheel visualizations display topics based on the volume of conversation behind that topic. That means if you search for Tesla you can guarantee that the word ‘Tesla’ will appear – a lot.
This doesn’t reveal anything new. Users almost always know the most talked about topics in their data.
Instead of focusing purely on the volume of topics within conversations, our topic wheel displays the topics that are more likely to be talked about by a given segment of the dataset.
To do this we let users segment data (by sentiment, gender, account type, etc.), letting Brandwatch identify topics based on how much they are associated with that segment. For example, instead of revealing that both men and women mention the ‘Tesla Roadster’, it will show that women talk more about its price, while men talk more about its acceleration.
These are specific words that over-index in the subset of data.
The topic wheel is available to all Brandwatch users at no extra cost. Simply log in and select the topic wheel from the component picker.
If you’re not a client but you’d like to see how your data looks, simply click the link below, enter your details, and we’ll organize a personalized demo.