Home » The Grasp #1: breaking AI, busting unions and binding books

The Grasp #1: breaking AI, busting unions and binding books

  • by
  • 3 min read

In The Grasp, I list three articles that deserve to be shared. The Grasp is about data in all its aspects. Articles from a wide variety of backgrounds will make it into the biweekly three. Some works will be technical, while others won’t. This format truly helps you grasp what’s going on in the fast-moving world of data. Curated by humans, no bullshit, and summarized to fit your schedule.

How COVID-19 broke AI

Machine learning is a process in which an algorithm is trained by looking at past data. During the pandemic, people started behaving in ways that the models had never seen before. Consequently, recommendation engines, sentiment analysis models, and inventory management algorithms went haywire.

“Machine-learning models are designed to respond to changes. But most are also fragile; they perform badly when input data differs too much from the data they were trained on. It is a mistake to assume you can set up an AI system and walk away…”

MIT Technology Review, Our weird behavior during the pandemic is messing with AI models (~5 min)

Union-busting big data

At Amazon-owned Whole Foods, employee behavior is quantified to estimate the probability that they will join a union. Besides “Team Member” sentiment, also store and external risks made it into the model that predicts in which of Whole Foods’ 510 stores, employees will organize in unions.

“Overall, US companies spent at least $100 million on consulting services for anti-union campaigns between 2014 and 2017, […] using a data-powered heat map to monitor for unionization risks ‘is just the next frontier of employer opposition to unions.‘”

Business Insider, Amazon-owned Whole Foods is quietly tracking its employees with a heat map tool that ranks which stores are most at risk of unionizing (~6 min)

Down the abyss of hand-drawn viz’

Edward Tufte named Charles Joseph Minard’s Carte Figurative “the best statistical graphic ever drawn.” An amateur historian of the Napoleonic era myself, I agree there’s probably no better way to visualize the human toll of the Russian campaign. In this piece, Paul Kahn traveled to Paris to revisit “The Minard System” of visualizing data.

“Minard would “strongly resist the tendency of the tyranny of precise geographical position to detract from the essential communication of his chosen theme.” To put it another way, Minard distorted geography when it allowed him to better tell the data’s story. His diagrams frequently include altered coastlines and changes to the relative size of countries, oceans and continents, always to accommodate the data.”

Paul Kahn (Medium), Touching Minard (~25 min)

Leave a Reply

Your email address will not be published. Required fields are marked *