Bradford Taylor

Data scientist at CDC. Tweets are my own. Interested in: social drivers of infectious disease, evolution of abx resistance, math models. He/him.

[Bradford Taylor @bradfordptaylor]

I worry the impact this paper will have on public health perception which is why scrutiny on model choice is necessary. Further, there needs to be more discussion on data quality since the results are from 2020-2021 when surveillance efforts were changing in time.

[Bradford Taylor @bradfordptaylor]

I'm not convinced--many models are just fitting noise. My understanding is, among the 100k models, 1/5 ask how deaths change 2 weeks after policy implementation. 2/5 for 4 week lag. The mortality time lag (~3 weeks) + ID dynamics means not enough time to see effect.

[Chirag Patel @chiragjp]

Haven't we learned anything about stopping pandemics? When it comes to government policy and COVID-19, the answer is: "no, we don't have the data."

I write an op-ed with Eran Bendavid in @statnews: https://t.co/HAtwxjajD2 about our research paper in @ScienceAdvances: https://t.co/lr4ftHXjMe

[Bradford Taylor @bradfordptaylor]

Kudos to their transparency: https://t.co/dQsTPHjaK7 . It's worth noting more models at the 4 week lag show beneficial effects, than the models with 2 week lag. I would've loved to see more sensitivity analysis of the model multiverse

[Bradford Taylor @bradfordptaylor]

@drugmonkeyblog If PIs should expect less from their trainees, should trainees expect more from their PIs? Per your blog:

[Bradford Taylor @bradfordptaylor]

@baym Something academics publicly scorn, but privately respect

[Bradford Taylor @bradfordptaylor]

@hsalis @ProfTomEllis So, akkermansia? Not sure I agree with that… obviously

[Bradford Taylor @bradfordptaylor]

@ClementYChow But the stated goal isn't to increase reproducibility, it's to aid communication, specifically during a talk

[Bradford Taylor @bradfordptaylor]

@baym Uh huh https://t.co/AimSMwjJnZ

[Bradford Taylor @bradfordptaylor]

@Celia_Sqe I've seen corresponding gmail addresses. Also, congrats!

[Bradford Taylor @bradfordptaylor]

@baym Clearly, this is why preprints matter

[Bradford Taylor @bradfordptaylor]

@wheezenfeld @_AlvinChristian Yeah, the *journal* club. Nerd!!

[Bradford Taylor @bradfordptaylor]

@reject_resubmit If these are indeed averages and not medians then the the skew/long tail of the distribution will lengthen the value, i.e., average becomes the long end. Bias in perception and not sample! Sincerely, a 7th year postdoc :-)

[Bradford Taylor @bradfordptaylor]

@neubadah I’m with you overall, but I disagree with the sentiment “coasting into” without “intention.” Moving somewhere new to take a low paying job signals intention to me regardless of competition for said job. Fewer PDs just means fewer PDs, not more committed PDs.

[Bradford Taylor @bradfordptaylor]

@baym Is that bone?

[Bradford Taylor @bradfordptaylor]

@wc_ratcliff What I find interesting is that we, by necessity, have operational definitions of these concepts, e.g., fitness = competition assay, species = ANI cutoff/gene flow, etc. With multiple definitions we can assess how fuzzy/robust these concepts are. Kinda empiricizes it

[Bradford Taylor @bradfordptaylor]

Are you an undergrad and love math modeling (or tolerate it and want to know more)? One week left to apply for the Workshop to Increase Diversity in Mathematical Modeling & Public Health! March 4-5, 2024 in Boston w/ @MIDAS_Network @CCDD_HSPH

https://t.co/I1PfpRANb4

[Bradford Taylor @bradfordptaylor]

@wc_ratcliff Really cool radio piece on the dynamics of social media spaces and why that "enshittification" might be inevitable: https://t.co/yPcabhWzyE

[Bradford Taylor @bradfordptaylor]

@318Glenn @HenryYin19 Surely, you JS’d

[Bradford Taylor @bradfordptaylor]

@TheSquishyLab Cool work! We found similar patterns and considered multiple infections in phage-bacteria spatial models, albeit using agent based models instead of PDEs: https://t.co/vcXiKoV2ux