Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Chelsea Parlett-Pelleriti Profile picture
Chelsea Parlett-Pelleriti
@ChelseaParlett
📊Statistician, 👩🏻‍🏫Data Science Prof, ✍️Stats Communication. #statistics #scicomm #datascience #statstiktok 👩🏻‍💻 she/her
2 subscribers
Apr 10 • 7 tweets • 3 min read
🪧Solution Thread🪧

We want a confidence interval for a two-sample z test! We're interested in the difference between replication rates for projects with and without open data.

This question has answer choices, so let's talk about my FAVORITE WAY to save time on these Qs!

2 sample proportion James is interested in whether projects on OSF (an open science platform) with open data (data that is available for download) are more replicable than projects without open data. He collects separate random samples of projects from OSF. Here are the results: (a 2x2 table is show with counts of replication and whether or not a paper had open data. The patterns show that open data papers tend to replicate more often). James wants to use these results to construct a 95% confidence interval to estimate the difference in the proportion of projects that reproduce (open - ...
Answer choices. The answer choices vary in their critical z value, and the way the standard error is calculated. The correct choice is A. FIRST: take a look at what's different between the answer choices. Here, the differences are:

- the critical z value
- the formula for calculating the standard error

so we don't even need to look at the other parts, they're all the same!
Aug 9, 2023 • 14 tweets • 6 min read
📚 A Few of My Favorite Data Science Related Textbooks

(I use the term textbook loosely)

Here are some (non-exhaustive) of my favorites: https://t.co/BCgfQadCmN
📕 Elements of Statistical Learning

This book was the BIBLE in Grad School. It’s incredibly in depth and dense, but not so much that you can’t get through it. It’s comprehensive, well written, and is my go to reference to understand a ML algo more deeply😍 The book “the Elements of Statistical Learning”
Jul 15, 2022 • 5 tweets • 1 min read
United healthcare’s student health insurance was a HUGE stressor during grad school🙃

I had to float large sums of money waiting for reimbursement (which they messed up ~40% of the time at first), spend hours on the phone with them, and had a hard time finding a therapist😤 But I’m glad they made SOOOOO MUCH money 🙄

Healthcare is a right, and shouldn’t be tied to school/employment. It should be accessible, largely free, and it should focus on helping people NOT MAKING PROFITS.

This stuff pisses me off.
Jul 31, 2021 • 11 tweets • 4 min read
Another #ChelseaExplains 🧵 (trying to start with simpler topics).

Today that's 💫Conjugate Priors💫 First, PRIORS. In Bayesian Statistics, we use probability distributions (like a normal, Cauchy, beta...) to represent uncertainty about the value of parameters.

Instead of choosing ONE number for a param, a distribution describes how likely a range of values are A distribution with the x-axis label "Possible Paramete
Jul 30, 2021 • 7 tweets • 3 min read
True to my word, the best statistical model, a Thread🧵

As an applied statistician and freelance consultant, I work a LOT with people trying to figure out the best model. Here are things I consider and ask. 1. Does the model ACTUALLY answer a question you have?

If you don’t have a question THEN THE BEST INFERENTIAL/PREDICTIVE MODEL IS NO MODEL. Do some EDA first!

Stop doing hypothesis testing if you don’t have a hypothesis to test. ✋
Jul 6, 2021 • 20 tweets • 7 min read
SINCE @kierisi has threatened to sarcastically/chaotically say incorrect things about p-values during #sliced tonight 😱 just to annoy people 😉,

I thought I’d do a quick thread on what a 🚨p-value🚨actually is.

🧵
(1/n) Computationally, a p-value is p(data as extreme as ours | null). Imagine a 🌎 where the null hypothesis is true (e.g. there is no difference in cat fur shininess for cats eating food A vs. food B for 2 weeks), and see how extreme your observed data would be in that 🌎 (2/n)
Nov 20, 2020 • 9 tweets • 6 min read
⚠️ SO YOU WANT TO BE A BAYESIAN⚠️ :

(since I compiled a quick list of Bayesian resources today, I figured I should share; these are just my opinion!) 📚 Beginner book (R): Bayesian Statistics the fun way —amazon.com/dp/B07J461Q2K/…

📚 Beginner-Intermediate book (R + JAGS + Stan): Doing Bayesian Data Analysis: amazon.com/Doing-Bayesian…
Nov 18, 2020 • 5 tweets • 2 min read
Wanna become a data scientist?

Sᴛᴇᴘ 1: ɢᴇᴛ ᴅᴀᴛᴀ.
Sᴛᴇᴘ 2: ᴅᴏ sᴄɪᴇɴᴄᴇ. Alternatively:

Step 1: Think about becoming a lawyer but ditch that because you can’t stand foreign political history classes. Add a philosophy double major because sureee that’ll help🙄. Then switch to psychology, meet an awesome statistician and decide you love statistics
Aug 14, 2020 • 4 tweets • 1 min read
Pangolins that look like grad students asking their advisor if there's funding for them for next year: A Thread🧵 "Should I apply for GRFP?"
Mar 4, 2020 • 8 tweets • 3 min read
Moving from psych to stats/DS is totally doable Depending on your training, there may be some gaps you need to fill, content-wise, but those gaps 1) aren't insurmountable + 2) will not automatically make you a bad data person just because you're working on filling them.

1/8 Doing good DS requires hard work/rigor but it’s not exclusive to “math” people. You can do it.

2/8