Data Scientist. Creator of https://xitter.cf/t.co/ByO92M3ceQ and https://xitter.cf/t.co/k9yZN1fORq. Presenter of unbiased takes. Realist. @MIT '15.
New York, NY
Joined February 2009
- Tweets3.7K
- Following614
- Followers64.4K
- Likes2.8K
Pinned Tweet
It's popular to use the US as a punching bag these days, but I can't think of many other countries where the work of an untrained 27-year-old immigrant can garner the attention & respect of researchers and public alike.
America isn't perfect, but I'm grateful to call it my home.
Global excess deaths in 2020-2021 is almost 3x higher than reported #COVID19 deaths. Great work by @WHO & @UNDESA to get this out. It's probably the most comprehensive estimates to date.
xitter.cf/WHO/status/15221959708…
The total number of deaths associated with #COVID19 worldwide from 2020-2021 may be closer to 14.9 million: New estimates by WHO & @UNDESA.
That’s 9.5 million more deaths than reported bit.ly/38SJIpM #HealthData
Happy Monday. I enjoyed my time on the Data Science Show with @DalianaLiu to talk data, Covid and lessons learned along the way.
If you want to get into data science, I highly recommend giving Daliana a follow.
youtube.com/watch?v=roRAhBeV…
Youyang Gu retweeted
My new post is out: ourworldindata.org/longtermi…
It is about humanity's past and future.
I don't know how to summarize this post in a thread. But I can share the two visuals I made for it. 👇
Income inequality is one of the best predictors of COVID-19 deaths in the US.
From @TheEconomist: "[Reducing inequality] can hardly be done overnight. In the meantime, governments need to tailor their pandemic response to take account of inequalities."
economist.com/finance-and-ec…
"In a recent exercise Youyang Gu, a data scientist, ran multiple versions of a model that seeks to find correlations between 41 variables & American state-level deaths from covid-19... [He finds] inequality has the biggest effect."
Here is that exercise:
xitter.cf/youyanggu/status/14074…
I believe income inequality is the single best predictor of total Covid deaths in the US.
Not income, but income *inequality*. The R^2 is surprisingly high: 0.35.
Thanks to @joe_sill, I was able to look at over 40 different variables, and expands on my analysis from last month.
There's a @nytopinion article regarding universal masking in schools being shared by many, including some prominent experts.
But the study never had a control group, so it's scientifically inaccurate to make any conclusions.
Here's a good critique of that piece:
THREAD
I corresponded with the authors of the Duke study behind this opinion piece weeks ago... nytimes.com/2021/08/10/opini…
This is another good example of the double standard in science and academia when it comes to studying pandemic interventions:
xitter.cf/youyanggu/status/14237…
Great interview with Prof @BallouxFrancois about Covid & scientific discourse. We need more pragmatic researchers like him.
"As a scientist, you need to be open to surprise and to let your priors be updated by new data. Being dogmatic is problematic."
theguardian.com/world/2021/a…
"Districts that went strictly remote experienced 42 percent more decline [in kindergarten enrollment].
City schools, which serve disproportionate numbers of low-income students of color, were the most likely to shutter classrooms for extended periods."
nytimes.com/2021/08/07/us/co…
People were worried cases would spike further when the UK lifted almost all remaining restrictions in July. The opposite happened.
We must acknowledge that restrictions aren't all that effective in Western countries.
(Except Australia, who just entered their 6th lockdown)
To clarify, I am not anti-lockdown. If lockdown/restrictions work and the benefits are worth the trade-offs, then I'm all for it. I just haven't seen solid evidence of causal effectiveness & a thorough discussion of the trade-offs.
I'm anti-bad science.
xitter.cf/youyanggu/status/14237…
As I didn't realize this thread would blow up, some additional clarifications:
1) My claim is based on my past analysis on the effectiveness of restrictions in the US: xitter.cf/youyanggu/status/13972…
2) Here is the original source which my Tweet was based from: businessinsider.com/uk-lifti…
The "younger & sicker" narrative for Covid doesn't appear to be supported by the data.
This is classic sampling bias: when the vax rate is much lower in younger populations, they'll unsurprisingly have more hospitalizations relative to older groups.
See: news.yahoo.com/statistics-sh…
One leaked CDC memo & headlines went from "99% of Covid cases are among the unvaccinated" to "majority of cases are among the fully vaccinated".
In reality, the truth is not very sensational: The vaccine is highly effective & working exactly as intended.
xitter.cf/youyanggu/status/14208…
By my estimates, Covid case rates among fully vaccinated individuals are currently ~4-7x lower than those that aren't fully vaccinated.
This is based on weekly data out of LA County. Unfortunately, our federal government has no real-time data on this.
xitter.cf/youyanggu/status/14176…