You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
A company found that listing high school lacrosse on one's resume correlates with high performance evaluations and therefore killed the project. One wonders what the expected happy don't-kill-the-project version of that analysis would have been. https://twitter.com/davegershgorn/status/1054357855678144514 …
"We took a data source which is basically known to be bullshit and then correlated it with another data source which we consider to be highly suspect with regards to fairness, on a naive bayesian model. We were sort of hoping this analysis would find nothing interesting." ?
I'm not following this. Are you saying they cancelled the project because they thought the ppl on the project would do an excellent job? So the project was never intended to succeed? (that doesn't surprise me but wanted to confirm)
I think my surprise is that, given the reason for canceling the project, there's basically no version of that project that can be successful, and indeed it's difficult to think of what success would look like for it even absent that reasoning.
Possibly "people who get frequent promotions tend to also have high performance evaluations". Rather than "resumes have zero predictive value", my takeaway from this story is "some people don't know how to properly discount the p-values of small-subset signals".
In my experience, "works at Stripe and is named Patrick" is a very strong indicator of awesomeness, but I would give it a very low weight in a predictive model because anything based on an N=2 sample needs a low weight.
Companies will keep learning this lesson: Resumes have nearly zero actual signal. ML isn't magic. It can't squeeze blood from a stone.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
