Dr David Greenwood

Data Scientist & Systems thinker @ Check-Risk: Using data science to reward risk-taking & enhance decision-making. Opinions expressed are solely my own.

[Dr David Greenwood @dsg22]

@IanSommerville Congratulations! Wishing you many more years of health & happiness 🙏 Thank you for the inspiration!

[Dr David Greenwood @dsg22]

Remember the 1929-1930 bull market?

The one that followed the 1929 crash?

when the Dow Industrials soared by 48%?

from 198.69 on 11/13/29 to 294.07 on 4/17/30?

then proceeded to collapse by 86% to 41.22 by 7/8/32?

Good times.

Be careful out there.

https://t.co/O4iYOtmOHa

[John P. Hussman, Ph.D. @hussmanjp]

Hey! Remember the 1929-1930 bull market?

You know, the one that followed the 1929 crash?

when the Dow Industrials soared by 48%?

from 198.69 on 11/13/29 to 294.07 on 4/17/30?

then proceeded to collapse by 86% to 41.22 by 7/8/32?

Good times.

Be careful out there.

[David Rosenberg @EconguyRosie]

Core retail sales are up while 14M jobs & $726B in wages have been lost. How? Personal income has RISEN at a 12% a.r. in 2020 (a $933B surge) — government transfers have ballooned at a 229% a.r. (by $2T), funded by the Fed. It may be “money” but it's not “income”. Big difference.

[John P. Hussman, Ph.D. @hussmanjp]

Hey! Remember the 1929-1930 bull market?

You know, the one that followed the 1929 crash?

when the Dow Industrials soared by 48%?

from 198.69 on 11/13/29 to 294.07 on 4/17/30?

then proceeded to collapse by 86% to 41.22 by 7/8/32?

Good times.

Be careful out there.

[Iain Begg @IainBeggLSE]

V, U, W or L - how will our economies emerge from Covid-19 and are there lessons from China? In collaboration with "QJ" from Fudan University, here is an analysis: https://t.co/JOH71H2Xto @LSEEI

[Andrew Gelman et al. @StatModeling]

Online Causal Inference Seminar starts next Tues! https://t.co/BlyO86mlM4

[Dr David Greenwood @dsg22]

Why do some crimes scale up faster than others? https://t.co/kPp3hlqWu3

[Dr David Greenwood @dsg22]

Did you think that Brexit or the election of Trump was 'not going to happen'? Our social networks create perception biases https://t.co/Euhr1CAhzn

[Dr David Greenwood @dsg22]

Signals suggest a correction may be coming soon...

[John P. Hussman, Ph.D. @hussmanjp]

I just ran this one for Felder's amusement.
@jessefelder

[Gabriel Peyré @gabrielpeyre]

Perron-Frobenius is fundamental for the study of Markov chains (convergence toward the stationary distribution) … and Sinkhorn’s algorithm (which is a non-linear extension). https://t.co/DiY6RRGJQ9

[Miquel Perelló Nieto @PerelloNieto]

Final arrangements being done for the 9th @PyDataBristol Meetup. We have reached the maximum of 100 interested to attend, but keep an eye in case of people unRSVP. #DataScience https://t.co/7gOOa6r4ln

[RedRock Consulting @weareredrock]

#TB to last month's @BristolRUsers meet-up at the RedRock offices - great talks, great food, great company.

Looking for an event space in #Bristol? We might be able to help. https://t.co/g1J5o98nPB #techevents #bristoltech #bristol

[Gabriel Peyré @gabrielpeyre]

Any optimization problem is equivalent to a convex (linear) one (but infinite dimensional…). The key do perform global optimization using Lasserre’s relaxation via the problem of moments (aka sum-of-square relaxation). https://t.co/L16MadLlFN

[Ben Hoyle @bjh_ip]

This is a great example from #bookofwhy showing why we need to be wary of taking the data as-is, while ignoring its generation.

Normal ML would “find” the relationship on the right.

[Bristol R User Group @BristolRUsers]

Just under 2 weeks now until the next #BristolR, and we have some excellent #rstats talks lined up! If you haven't signed up yet, why not? 😀 https://t.co/9rcpnTgpye

[John P. Hussman, Ph.D. @hussmanjp]

Rare for leadership and participation to be so ragged, so close to record highs, as they were today. Typically not a great sign.

[David Berthelot @D_Berthelot_ML]

MixMatch (https://t.co/kfMnrJ4jBN) code is released https://t.co/M4Yx2mU2lp (Python3, TensorFlow 1.1x). Let me know how it works for you and also let me know if you port it to other frameworks.

[John P. Hussman, Ph.D. @hussmanjp]

This may be a good moment to note that the S&P 500 is only -4.4% below its steepest historical extreme. On one hand, that indicates this decline isn't material from a long-term perspective. On the other hand, the fact that it's not material means stocks remain hypervalued.

[Ernesto Carrella @ErnestoCarrella]

Estimating ABM parameters from data got you down? Is the literature on likelihood-free methods too big to know where to start? Would you prefer somebody else go through a bunch of estimation algorithms and check which one is best?
Lucky you! https://t.co/52pCSykECi
(1/2)

[Gabriel Peyré @gabrielpeyre]

Heat equation vs wave equation inside a planar domain. Boundary conditions matter for PDEs. https://t.co/7zVsdabTYt