The last 50 weeks in the world of data
Round up on the top DataScan stories so far.
Data is giving rise to a new economy — but unlike any previous resource, there is a “disincentive to trade” as data can be duplicated and used for other purposes. But if data “remains stuck in silos”, most of the value “may never get extracted”.
Information gets more valuable when it is shared, to the extent that disconnected data (due to silos and ineffective collaboration) costs organisations $140 billion a year.
Having a competitive advantage is now dependent on accessing huge stockpiles of data (alongside having strong consumer data protection).
Companies need a solid foundation of high-quality data to make their AI effective. If AI is fed garbage, the output will be the same → machine learning will “only ever be as good as the data”.
Regulation
The tech sector is struggling to prepare for the GDPR. Centralised databanks are too vulnerable to breaches, even with the most robust cyber-security.
The growth of decentralised tech is inevitable → has it already outdated the GDPR? Matthew Hodgson explains how a decentralised web can give people power back.
Researchers revealed that the re-identifying of anonymised data is “trivial” — read how it’s done here. Anonymous data-sharing networks advance medical research but are vulnerable to being cross-referenced to identify individuals.
Re-identifying data is due to be criminalised under the Data Protection Bill, but the actual crime should be the releasing of badly anonymised data.
Privacy
Why privacy ≠ secrecy:
Don’t confuse privacy with secrecy. I know what you do in the bathroom, but you still close the door. That’s because you want privacy, not secrecy.
Is America becoming Orwell’s nightmare? How Cambridge Analytica used big data and psychometrics to assist Trump’s campaign — to the point where “pretty much every message that Trump put out was data-driven”.
The Royal Free hospital broke the law when it gifted 1.6 million patient records to DeepMind. Elizabeth Denham: “it’s not a choice between privacy and innovation” — it needs to be done correctly and transparently.
The ICO fined 11 big UK charities for swapping and selling donor lists. The charities were “wealth screening” supporters and identifying new ones by piecing together personal info from other sources. A number of charities also created a massive pool of donor data for sale.
Data viz
Graphic designer Herwig Scherabon visualised the data behind gentrification. The Joy Division-esque diagram above shows the differences in rent prices across London (image above).
Sign up receive my weekly curation on the most important data news. 💌
