The official twitter account for https://xitter.cf/t.co/Smhdl3aYG0, by the @adamcchang lab.
Singapore
Joined April 2018
- Tweets150
- Following9
- Followers604
- Likes132
Estimation Stats retweeted
Nearly two weeks into the 'circuit-breaker' shutdown, Singapore's R remains well above 1. The exponential increase is largely due to infections among work-pass (WP) holders who are housed in crowded conditions.
Datavis by @jacuzzijo adapting @kevin's notebook.
Estimation Stats retweeted
A step in the right direction, but a better way is to meta-analyse the set of replicated experiments. I used DABEST-Python to plot each expt's effect size, and added a summary diamond effect size as seen in forest plots!
1/4
github.com/ACCLAB/DABEST-pyt…
@samjlord @FritzLaylin
If N=every cell observed, P can be <0.05 even with control data sets. Instead, we (@samjlord @FritzLaylin @MullinsLab) suggest showing reproduciblity between biological reps (triangles), while still displaying cell variability (dots) using SuperPlots!
arxiv.org/abs/1911.03509
Estimation Stats retweeted
"Estimation does not cure all ills, but this new policy can serve a vital role in #eNeuro’s forward-looking efforts to promote strong rigor and #reproducibility without sacrificing scientific vitality."
How @EstimationStats improves the research cycle: bit.ly/31bBMGW
Estimation Stats retweeted
The easiest way to never worry about power, false positives, or preregistration is to never make any decision as you do science. Never choose to do a study or abandon a research line based on a previous study. Added benefit: If you never make a decision you'll never be wrong :)
Estimation Stats retweeted
Really digging the 'dabestr' R package from @EstimationStats . It plots bootstrapped confidence intervals for effect sizes, encouraging quantitative rather than dichotomous evaluation of effects.
Estimation Stats retweeted
Had a go with @EstimationStats for a paper I’m writing & replaced ~200 words in results with ‘see fig 1’ 🥳👏🏼👏🏼👏🏼
Estimation Stats retweeted
Our latest work now out in @naturemethods! "Moving Beyond P Values: data analysis with estimation graphics". We show how estimation plots can improve statistical reasoning with elegant and informative estimation plots.
nature.com/articles/s41592-0…
1/3
Estimation Stats retweeted
Excellent free analysis and access to free visualisation packages also. A huge boon in many fields.
Moving beyond P values: Everyday data analysis with estimation plots biorxiv.org/content/10.1101/…
#Rstats package: github.com/ACCLAB/dabestr
Python package: github.com/ACCLAB/DABEST-Pyt…
Web: estimationstats.com/#/
Estimation Stats retweeted
The new version of our estimationstats.com preprint covers more topics, including larger samples, paired data and multiple groups. It's release coincides with @jacuzzijo's upgrades to the app, which now plots standardized effect sizes!
biorxiv.org/content/biorxiv/…
Estimation Stats retweeted
Please retweet: Blake McShane, Sander Greenland and I have been commissioned by Nature to write a comment on statistical significance and dichotomania. Wanna be co-signatory? For more info, send an E-mail until March 7 at [email protected]
Estimation Stats retweeted
dabestr, our #rstats package for #estimationplots, is now available on CRAN!
Estimation Stats retweeted
We are proud to announce dabestr, an R package for estimation stats! It allows fast and intuitive creation of estimation plots. We look forward to all your beautiful plots! #rstats #estimationstats #estimationplots
github.com/ACCLAB/dabestr
Estimation Stats retweeted
DABEST v0.1.4 is now available via Github github.com/ACCLAB/DABEST-pyt… and PyPi pypi.org/project/dabest/.
Features tweaked aesthetics alongside updated docs! acclab.github.io/DABEST-pyth…
Estimation Stats retweeted
Grouped-data plots (bars, boxes, dots, swarms, or violins) diagram the sample distribution. Estimation plots add a second axis to visualize the effect-size distribution, typically the distribution of the difference of means: ∆.
Violin plot ≠ ∆ curve.
Estimation Stats retweeted
Stats Quiz: The average NFL player is about 25 years old, just over 6'2" in height, weighs a little more than 244lbs and makes slightly less than $1.5M in salary per year. Can you tell which distribution goes with which trait? See why the shape of the distribution matters a lot?
