Catching up continues in 2019! Have a look at this completely subjective roundup of digital news.
Back to the future
It’s that time of year where new year’s wishes, look-backs, predictions, and speculations are all abound (not to mention those good resolutions for a healthier life… impossible to keep…). In its 10 social media trends for 2019, Kantar Media predicts notably that platforms will move “beyond algorithms”. Only 1 in 3 people consider social platforms to be reliable sources of information. From fake news to filter bubbles to data protection breaches, social platforms’ algorithms have hit their limits. What’s missing from these platforms? A dose of “human” intelligence for more refined interpretation of the multi-faceted consumption of content. Then again, that’s to be expected – these platforms were built on hypergrowth models. Science without conscience…
The ABCs of data literacy
The Data Literacy Project is a global initiative to promote data literacy; its goal is to help all companies’ employees master data. The project aims to improve people’s skills so that they can critically question data, take decisions, and share data insights. Today, only 1 in 3 people can understand data, analyse and question it. This project is an ambitious one that could spread beyond the professional sphere to fight against the proliferation of fake news, for example. When amplified by a social network, false or modified information can have devastating results. In India, false rumours spread on WhatsApp resulted in the death of several dozens of people. It’s therefore urgent to educate citizens so that they can discern truth in the avalanche of data they ingest.
After months of negotiations, the European Commission doesn’t have much to show regarding the taxing of the web giants – the half-hearted measures have been adopted pretty much by France alone. And our European partners’ cold feet means the piece of the tax pie will be considerably smaller. Lack of political courage? Or rather a show of force of tax-evading economic powers? Even beyond the fiscal aspects, monopolistic power is disturbing. Especially so for media agencies who, seeing the imminent danger, are rising up against opaque practices, lack of impartiality, and the closed ecosystems of “global” platforms which are now direct competitors. A major point of friction: the reliability of campaign results indicated by the platforms. Following the data restrictions that Google imposes (ad inventory and non-GDPR-compatible user IDs), agencies are worried about both the future of tracking and the reliability of the data provided. These agencies, rallying behind Gautier Picquet, the president of Publicis Medias, are calling for shared rules to be applied in the ad market regarding transparent measurement and access to data. On the advertiser side, good intentions have been shown (certification and reference measures), but the domination of platforms is still clear: 79% market share goes to online ads… Nonetheless, these business models are not unshakeable and could change. Why? Because it’s never too late.
Diapers and beer
An algorithm for extracting frequent correlations (or association rules) famously discovered that customers who purchase beer also buy… diapers. The sociological explanation for this surprising association? New fathers, (temporarily) in charge of the family, take the opportunity to restock their fridges with beer while picking up diapers for the baby. This example of algorithmic application is striking because it’s truly unexpected. Learn about the origins of this now-legendary metaphor in the data mining world, and understand how association rules in data mining work in in this article.
See you next time on the Internets!
- Back to the future. www.comicbook.com