“The rate at which we’re generating data is rapidly outpacing our ability to analyze it”
These were the words uttered by Patrick Wolfe, Executive Director of Big Data Institute at the University College of London, to Business Insider. It captures the current state and hype around Big Data. An estimated 2.5 quintillion bytes of data is being created every single day. The capabilities to analyze all this information has yet to be developed.
The power of predictive analytics
Wolfe goes on to say that only 0.5% of data being generated is analyzed. These words resonate quite strongly in the face of recent events. If you consider two recent landmark events on a global scale, it would be the election of Donald Trump and Britain’s’ exit from the European Union. Neither of the events were predicted convincingly enough by broadcasters, analysts and everyday citizens.
Trump’s election was predicted by South African company BrandsEye
In fact, the result came as a surprise and shock to everyone following these events. From a predictive analytics point of view, it is either that correct data was not being analyzed or no analytics were carried out at all.
There was however a company that predicted both events spot on. They predicted Trump would be elected and that Britain would leave the European Union. Even more fascinating, they were not based in either of those countries, or even on those continents. A company called BrandsEye in South Africa accurately predicted both of these events even when other famous large scale newspapers and broadcasters got it all wrong- but, how?
It was the careful way and integrity with which they treated their data. An example of this is with the prediction of the Brexit. BrandsEye studied over 10 000 tweets, which represented about a half a million opinions. They registered the sentiment of each Tweet as to whether it was positive or negative. The outcome was that they predicted a 56.9% chance for the exit of Britain from the European Union compared to the 47% shown by newspaper polls. They did two things right: they got social media feeds from an accurate sample according to demographics and size. They also took time to read through each tweet and register any sentiment of sarcasm or irony, which can often disturb the integrity of the data.
BrandsEye used Sentiment analysis to predict the Brexit
It’s an extremely costly and time-consuming process because automated processes have yet advanced. But, it’s all the more rewarding. Analyzing social media to predict the election of Trump and the Brexit is pretty spectacular – all these predictions made from another continent. Reading into the future almost sounds like something from a sci-fi movie – but it’s very real and possible as shown.
The predicament of predictive analytics in an African context
Using similar techniques, the South African local elections are being predicted and seeing where it will go and the reactions it will cause. The integrity of the outcome might be valid because a large population of South Africa uses social media thereby making it easy to predict. However, the huge limitation from an African perspective is that only 9.1% of Africans use social media. To use any type of predictive analytics with social media would be unethical and quite honestly foolish at this point.
Consequently Kenya has moved towards improving polling systems and stepping away from predictive analytics for now. An app called “Vote Now” is being used instead as a tool to understand how voters will likely choose their next government in the Kenyan election happening August 8. Due to limitations, predictive analytics is being treated with skepticism instead a medium with great potential.
It’s quite a challenging situation to find yourself in. Africa presents itself as a continent of two tales. On the one hand, recent news have shown DotModus, a South African based data company, reaching the Google Cloud Partner Specialization status – a big stamp on the data capability from an Africa perspective. On the other hand, you have Guinea Bissau that has an internet penetration of only 4.3%. The unfortunate part is that global partners are investing in locations that are already developing and showing economic prosperity such as Nigeria, Kenya, and Uganda.
Often the smaller areas are ignored leading to the gap in social media capability widening on the African continent. Companies such as MTN and Huawei have recently opened up big Data labs in South Africa and Nigeria.
The big question is, what is happening with the rest of the 52 countries on the African continent?