Analysis of #FeesMustFall tweets reveals startling trends

Share on facebook
Share on twitter
Share on linkedin
Share on email

Cyber security firm Snode has taken a look at tweets sorrounding recent #FeesMustFall protests and found some incredibly interesting data.

Earlier this year Snode launched a cyber security solution which uses machine learning and algorithms to detect patterns and anomalies in a network.

We’ve learned that Snode is quite good at detecting patterns so the team decided to see what sort of patterns they could find while analysing tweets related to #FeesMustFall.

This was done by looking at a few things namely; location of tweets and tweets using the #FeesMustFall hashtag.

So what did they find?

Looking at the locations of tweets Snode found that many tweets about Wits University originated from Pretoria. This Snode says contrasted against the subject of the tweets analysed.

Snode also found that users in Pretoria referenced Wits University 14 times more than the University of Pretoria.

Another interesting revelation was that of those users in the capital only 3% of #FeesMustFall tweets came from users linked to the Wits University account.

The vast majority of tweets referencing the institution – 94% to be precise – were sent from accounts that have political ties.

This, says Snode reveals that the #FeesMustFall protests may have been adopted by social media users to draw attention away from the goals of students.

Data in pictures

Snode says that by further analysing the meta-data in photos shared on Twitter there exists the potential to glean even more data such as the location the photo was taken (using GPS co-ordinates), the type of device the person was using and even a user’s real name.

But it doesn’t end there according to Snode co-founder Nithen Naidoo.

“There are a host of AI applications available, such as Russia’s FindFace, which allows users to scan a digital image of someone and then discover their online profile. There are therefore a number of tools on hand to benefit law enforcement as much as they do criminals,” said Naidoo.

This analysis shows us that you can’t inherently trust everything you read on social media, and perhaps we should be more questioning of what we see rather than sharing something because we think it’s topical.

Deep analysis of social media can also help those in positions of power make more informed decisions about what the public sentiment really is.

“With regard to socially relevant topics such as the student protests, having cyber intelligence at work can assist in gleaning vital insight,” says the co-founder.

“At Snode, we believe having such knowledge can not only help us understand the climate better, but also assist law enforcement and government services to predict and respond to critical events more efficiently,” Naidoo concluded.

[Image – CC BY SA Esther Vargas]

Brendyn Lotz

Brendyn Lotz

Brendyn Lotz writes news, reviews, and opinion pieces for Hypertext. His interests include SMEs, innovation on the African continent, cybersecurity, blockchain, games, geek culture and YouTube.