Share this page!

Philippe Lemoine Profile picture
Twitter logo
23 May, 6 tweets, 2 min read
At least I didn't suffer in vain. Image
Getting Rcpp sugar to work with data that were created with RcppArmadillo is a huge pain, though hopefully it's just because I had never used Rcpp before and it will be easier next time, but it's worth it.
I hadn't realized that R was so fucking slow with loops. I guess now every time I have a routine using loops that can't easily be vectorized, and I need it for computationally intensive tasks, I'm writing that little bastard in C right away.
I just need to figure out how to distribute the routine in a package so I can use it with future_map, but that can't be very hard. (Famous last word.)
Science is still going to get owned, but now it's going to happen 10 times faster 👌
That's what I will tell my grandchildren on my death bed 😪

• • •

Missing some Tweet in this thread? You can try to force a refresh
　

Keep Current with Philippe Lemoine

Philippe Lemoine Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @phl43

Philippe Lemoine Profile picture
24 May
It's been really painful to watch the reactions to this story, which as this thread notes has all the usual red flags (people never learn), but the worse part may be that, depending on the details, *even if the story is true*, it could be totally uninteresting.
As I noted previously, more than 1,000 people work at the WIV, so even if 3 people got symptoms compatible with COVID-19 in November, that would be less than 0.3% of the staff. Would that be surprising? Well, it depends on details the story doesn't give. necpluribusimpar.net/did-china-lie-…
In the US, the CDC estimates that ~8% of the population gets sick because they are infected by the flu each year on average. For influenza-like illnesses (ILI) in general it must be significantly more. It's probably about the same in China. cdc.gov/flu/about/keyf…
Read 41 tweets
Philippe Lemoine Profile picture
21 Apr
I've heard this objection over and over again in response to my argument that lockdowns don't pass a cost-benefit test and it's such an obvious non-sequitur that it just baffles me that people keep raising it. marginalrevolution.com/marginalrevolu… Image
The idea is that, if lockdowns don't make a big difference because voluntary behavior changes have a similar effect on transmission, they also don't make a big difference on people's well-being because people are going to do the same thing no matter what.
But that's a non-sequitur because the fact that state-enforced lockdowns don't have a large effect on transmission beyond what voluntary behavior changes would achieve in the absence of government interventions doesn't mean that people behave in the same way under a lockdown.
Read 5 tweets
Philippe Lemoine Profile picture
17 Apr
I'm very happy to announce that my paper has been accepted for publication by the European Journal of Epidemiology.
Folks, before you take this seriously (although in a sense it’s very serious), I recommend that you actually read the abstract 😄
The worst part is that I could totally write that paper. The modeling itself would only take 30 minutes, but the really fun part would be the write-up.
Read 4 tweets
Philippe Lemoine Profile picture
16 Apr
Sauf qu'il n'y a aucune raison de penser que E(années de vie restantes | age = x & sexe = y & victime du COVID-19) est égal à E(années de vie restantes | age = x & sexe = y) et qu'il est même parfaitement évident que ce n'est pas le cas 🤷‍♂️
À n'importe quel âge, l'immense majorité des gens qui sont infectés par SARS-CoV-2 survivent, donc ceux qui en meurent sont vraisemblablement plus fragiles que la moyenne des gens du même age et du même sexe et auraient sans doute vécu moins longtemps.
Ce tableau est donc trompeur dans le contexte du débat sur le nombre d'années de vie perdues par les victimes du COVID-19. Bref, avant de faire le mariole et de donner des leçons de démographie aux autres, mieux vaut réfléchir un peu et s'assurer qu'on ne dit pas de connerie...
Read 6 tweets
Philippe Lemoine Profile picture
16 Apr
Not only did your study show no such thing, but it rests on demonstrably false assumptions, so it's really extraordinary that you continue to peddle those results. Here is a thread in which I explain why this study is worthless and should never have been used to guide policy 🧵
First, the model used in that study assumes that B.1.1.7, the UK variant, is 59% more transmissible than the historical lineage. This estimate is based from Gaymard et al. (2021), which obtained it by fitting a simple exponential growth model to only 2 data points from January.
As I explained at length in this post, even if we just use those 2 data points from January, this estimate is highly sensitive to the assumptions we make about the distribution of the generation time and there is a lot of uncertainty about that. cspicenter.org/blog/waronscie…
Read 16 tweets
Philippe Lemoine Profile picture
14 Apr
Des nouvelles de B.1.1.7, le « variant anglais » qui était censé provoquer un tsunami en raison de sa transmissibilité accrue, à partir des dernières données de Santé publique France 😂 Image
Même chose mais quand on fait la comparaison uniquement avec la souche historique plutôt qu’avec tous les variants non-B.1.1.7. En gros, la première méthode est sans doute un peu biaisée, tandis que celle-ci ne l’est pas mais l’erreur de mesure est plus grande. Image
Je rappelle que les génies de l’Inserm et de l’Institut Pasteur continuent de faire l’hypothèse qu’il est 50% à 70% plus transmissible dans les modèles qu’ils utilisent pour faire les projections qu’ils présentent au gouvernement 👌
Read 5 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!

Like this author's thread?

Get emails when @phl43 has new unrolls!

Check out other threads from @phl43!

Read all threads by @phl43

:(