Allen Downey @AllenDowney Twitter profile

Pinned Tweet

Allen Downey

7 months

My copy of Probably Overthinking It has arrived! It should be shipping soon. If you would like to pre-order, you can get 30% from University of Chicago Press. Use the code UCPNEW.

14

10

139

Last Seen Profiles

@wizeguy205

@Vikasmalhotraaa

@TheGDeetz

@PreciousAyesiga

@WxRab

@heyimghouls

@FirdausiAhmad14

@Law_Marketing

@bbw_x_

@pretty_boy_dan

@transrimmermans

@madrid5681

@0IV6zWbI2bXsPdL

@stw_pdg

@DPFPnews

@SArrowsEsports

@AJG_DFS

@MUWalumniassoc

@NeoLin7812

@Therealmadmoney

@Ross2Raww

@Maple_CANE

@brody_goc

@lmbs45

@bagsfan75_Tony

@GandGFitnessEq

@Uf0AGGQhl8wHKxB

@Michell04221528

@ShunkoDaii

@Hyemathkashmir

@ISAM198002

@AlvaradoJsus

@trick4pa

@atomicmassgames

@pncrkesepian

@TajWazi01540136

Allen Downey

@AllenDowney

1 year

Today I asked ChatGPT to solve almost every exercise in Think Python and DSIRP. It did. My conclusion: everyone who writes code should spend the next month doing professional development on writing code with LLM-assist. This is how code will be written from now on.

156

266

2K

Allen Downey

@AllenDowney

1 year

In the last week, three people on reddit/r/statistics have asked about testing whether a sample came from a Gaussian distribution. The answer is that you should never test for normality. The result is a non-answer to the wrong question.

Never Test for Normality

Way back in 2013, I wrote this blog post explaining why you should never use a statistical test to check whether a sample came from a Gaussian distribution. I argued that data from the real world n…

www.allendowney.com

38

218

2K

Allen Downey

@AllenDowney

2 years

This semester I wrote a book, Data Structures and Information Retrieval in Python. It covers data structures and algorithms, organized around a motivating example: building a search engine. It's all in Jupyter notebooks that run on Colab. With quizzes!

Data Structures and Information Retrieval in Python

I am happy to announce the first release of a new book, Data Structures and Information Retrieval in Python, which is an introduction to data structures organized around a motivating example: build…

www.allendowney.com

15

248

1K

Allen Downey

@AllenDowney

3 years

Hey, look what I got!

16

42

1K

Allen Downey

@AllenDowney

6 years

A few years ago I wrote a short a book that explains basic use of Git: It contains exercises you can do on a practice repository: Today I added a page called "Merge conflicts with minimal pain": Good luck!

GitHub - AllenDowney/blair-walden-project: The Blair Walden Project: in 1845 Henry David Thoreau...

The Blair Walden Project: in 1845 Henry David Thoreau went to live in the woods... a year later his journal was found. - AllenDowney/blair-walden-project

github.com

7

305

995

Allen Downey

@AllenDowney

1 year

Math notation is good for a lot of things, but representing algorithms is not one of them. Fortunately, we have other formal languages that are really good at representing algorithms -- programming languages. It's a shame academic papers don't use them more often.

30

69

974

Allen Downey

@AllenDowney

2 years

Modeling and Simulation in Python is now available for preorder, and Barnes and Noble has it on sale for 25% off:

15

98

746

Allen Downey

@AllenDowney

4 years

The "Elements of Data Science" curriculum is now substantially complete: It is a notebook-based introduction to Data Science in Python for people with no prior experience in programming or statistics.

9

209

711

Allen Downey

@AllenDowney

4 years

I have started work on the second edition of Think Bayes. I think it will be much better! Here's the notebook for Chapter 1: There are exercises at the end if you want to play along at home.

Google Colab Notebook

Run, share, and edit Python notebooks

colab.research.google.com

9

128

678

Allen Downey

@AllenDowney

17 days

The 3rd Edition of Think Python is available now at The print edition is available for preorder, expected to ship in June. What's new? Jupyter notebooks, turtle graphics, doctest, unittest, regular expressions, and a new, full color, parrot on the cover!

11

164

658

Allen Downey

@AllenDowney

4 months

This is interesting if true, but I'm not sure it is. I tried to replicate the US graph with GSS data and I'm not seeing it. Whether the gap is growing depends how seriously we take the last two points is a noisy series. But it's nowhere near 30 points.

John Burn-Murdoch

@jburnmurdoch

4 months

NEW: an ideological divide is emerging between young men and women in many countries around the world. I think this one of the most important social trends unfolding today, and provides the answer to several puzzles.

2K

15K

56K

24

114

645

Allen Downey

@AllenDowney

4 months

A recent article in the Financial Times claims that there is an increasing ideological gender gap in several countries. In this article, I replicate their analysis with GSS data and conclude that there is little evidence the gap in the US is growing.

18

90

559

Allen Downey

@AllenDowney

2 years

A modest proposal: Let's stop using the term "bias-variance tradeoff". "Underfitting" and "overfitting" are clear, self-explanatory, and easy to remember and recognize. "Bias" and "variance" add nothing but confusion. And the word "bias" is already too overloaded.

31

52

553

Allen Downey

@AllenDowney

1 year

Modeling and Simulation in Python is off to the printer! And available for pre-order :) To celebrate, I published on of my favorite examples: One Queue or Two?

8

62

505

Allen Downey

@AllenDowney

3 years

In summary: Python Python Python Python Python Python

Randy Olson

@randal_olson

3 years

What #programming languages are required for data jobs at @Meta . #DataScience #MachineLearning #dataviz Source:

52

428

2K

9

76

494

Allen Downey

@AllenDowney

3 years

#1 New Release in Statistics...

6

37

485

Allen Downey

@AllenDowney

5 years

New rule: after I explain something, if a student takes a photo of the board, I get to be in it.

6

43

485

Allen Downey

@AllenDowney

3 years

I have news! After 18 years at @olincollege , I am leaving at the end of this academic year. Since I turn 55 in May, I am retiring, technically :) Not sure yet what I'll do next. I'll take some time to figure it out -- watch this space!

37

8

440

Allen Downey

@AllenDowney

2 years

Planning my Data Science class for the spring, I have 7 slots reserved to critique interesting data visualizations: what do you notice? what do you wonder? what works? what would you try changing? The NYT cochlea of COVID will be on the list. What else?

45

40

432

Allen Downey

@AllenDowney

5 years

1952: Persecution 2009: Apology 2013: Royal pardon 2019: Alan Turing on the 50 pound note.

8

105

397

Allen Downey

@AllenDowney

5 years

Most of the teaching examples for classification algorithms are toy data, fake applications (looking at you, irises and Titanic datasets). Any suggestions for examples with real data, real applications, ideally in the space of data science for social good?

45

65

393

Allen Downey

@AllenDowney

3 years

@causalinf If grandfathers count: when I was 10 I got to stay up all night in the print room of the Boston Herald. I got (and still have) my name in several typefaces on metal slugs. And I got to press the big, red STOP THE PRESSES button! Which is actually big and red.

5

3

381

Allen Downey

@AllenDowney

6 years

The second edition of Modeling and Simulation in Python is done: I am printing copies for my class this fall. If you would like a hard copy, you can get one from Lulu: Cover design by Olin's own Tim Sauder.

6

117

377

Allen Downey

@AllenDowney

6 years

Can someone explain why, if you write an idea in math notation, that's "theory", which provides deep understanding of the math "behind" it, but if you write the same idea in a programming language, it's just hacking? This bizarre prejudice is the bane of my professional life.

29

78

353

Allen Downey

@AllenDowney

2 years

Programming languages like Python are more readable than pseudocode, and have the additional advantage of being executable and debuggable. Pseudocode is obsolete.

Julyan Arbel

@JulyanArbel

2 years

Seek & Find: Will you spot the 3 errors hidden in this algorithm? Published paper, true story. Apart from that algorithm, I liked the paper :-)

16

9

97

16

36

329

Allen Downey

@AllenDowney

1 year

We have a cover! What do you think?

21

12

319

Allen Downey

@AllenDowney

3 years

I know I shouldn't always take the bait, but someone on the Internet was mean about Bayesian statistics, so I wrote a manifesto: "Bayesian and frequentist results are not the same, ever"

Bayesian and frequentist results are not the same, ever

I often hear people say that the results from Bayesian methods are the same as the results from frequentist methods, at least under certain conditions. And sometimes it even comes from people who u…

www.allendowney.com

12

57

308

Allen Downey

@AllenDowney

2 years

In my academic career, I have taught about 80 classes for a total of ~2000 class meetings. Today at 4pm will be my last. (At least for a while)

17

3

301

Allen Downey

@AllenDowney

5 years

Got a helpful email today from a student reading my book with a screenreader. Among the suggestions for better accessibility: use "can not" rather than "can't"; with a screenreader it is hard to hear the difference between "can" and "can't". And that's an important difference!

7

43

287

Allen Downey

@AllenDowney

6 years

Why learning to program is actually getting harder, and what we can do about it:

23

132

287

Allen Downey

@AllenDowney

4 years

I've been working on a new book about SQL, Astropy, Pandas, and Matplotlib. Here's the first public draft (generated with Jupyter Book)

5

52

279

Allen Downey

@AllenDowney

5 years

Think Julia will be available soon at an Amazon near you: To my surprise, O'Reilly has gone with a radically new cover design!

2

49

259

Allen Downey

@AllenDowney

5 years

For my Data Science class in the spring I am compiling resources related to data visualization. What books would you recommend? Web sites? Other? I will collect suggestions and post them next week.

56

44

254

Allen Downey

@AllenDowney

1 year

Ruin this joke by explaining it? Ok! This is an instance of Berkson's paradox. If a dingy restaurant with a broken website has bad food, it won't last, so among surviving restaurants, greasy styrofoam is correlated with good food. There's a chapter about this in my book!

Dan Wilbur

@DanWilbur

1 year

I was checking to see if a burrito place near me had good reviews and the photos of the place look dingy and the food looks greasy and they serve it in styrofoam and when I clicked their website I got a 404 error. This is going to be the best food I’ve ever tasted.

6

73

1K

8

29

243

Allen Downey

@AllenDowney

8 months

The Overton Paradox in three graphs 1. Older people are more likely to say they are conservative.

7

45

238

Allen Downey

@AllenDowney

3 years

I'm happy to report that the second edition of Think Bayes is available for preorder now. The red striped mullet says "order yours now!"

3

31

238

Allen Downey

@AllenDowney

5 years

Scientists: if you are still writing articles primarily in the passive voice because you think journals require it, please check the style guides. Many journals, include Nature, Science, and PNAS, have been begging you to stop for years.

3

82

221

Allen Downey

@AllenDowney

4 years

I'm working on a new series of notebooks to teach probability and Bayesian statistics. The first notebook starts with a famous example of the conjunction fallacy, Tversky and Kahneman's Linda the banker. Want to play along at home?

01_linda.ipynb

Run, share, and edit Python notebooks

colab.research.google.com

6

44

217

Allen Downey

@AllenDowney

5 years

Today is my first day on an exciting new project. This semester I will be at Harvard one day a week, co-teaching a seminar on data science education and helping to design a new undergrad class for Spring 2020. Details at

New Data Science Course!

STAT 250: Graduate Seminar on Designing the Elements of Data Science Course Fall 2019, Mondays, 3-5:30 pm, Science Center 706 Website: stat250.org Learn data science by teaching it! Data science co…

www.hsph.harvard.edu

11

20

213

Allen Downey

@AllenDowney

6 years

Hot off the press!

0

33

207

Allen Downey

@AllenDowney

2 years

Anybody have a favorite introduction to SQL (book or online resource) for a student who wants to do an independent study?

44

35

205

Allen Downey

@AllenDowney

3 years

More examples of Simpson's paradox in the General Social Survey. Old people are racist, sexist, and homophobic, but it's not because they're old; it's because they were born a long time ago.

7

34

207

Allen Downey

@AllenDowney

5 years

In my Complexity Science class, I mentioned the way NumPy creates "views" to avoid copying arrays. To explain the idea more clearly, I created this notebook, which you can run on Colab: It has some exercises you can work on, in case you are bored.

views.ipynb

Run, share, and edit Python notebooks

colab.research.google.com

2

46

202

Allen Downey

@AllenDowney

3 years

The slides, video, and notebooks from my Survival Analysis tutorial are available now from

4

35

202

Allen Downey

@AllenDowney

3 years

If you've been vaccinated, thank a scientist. And then thank about 100 engineers. Because inventing mRNA vaccines is science. But manufacturing billions of doses, keeping them cold, and delivering them around the world is a feat of engineering.

5

24

203

Allen Downey

@AllenDowney

3 years

I've been working on Modeling and Simulation in Python for... a while. On my 4th try, I have a version I am happy with. It's still a work in progress, but I've posted a mostly complete draft:

3

37

200

Allen Downey

@AllenDowney

4 years

I don't understand why @github STILL can't reliably render a Jupyter notebook. It's been years! NBViewer is 100% reliable as far as I can tell. Why is this so hard?

13

14

195

Allen Downey

@AllenDowney

1 month

Which plot indicates a stronger relationship? Discussion here:

18

25

198

Allen Downey

@AllenDowney

3 years

Metaphorical use of "growing exponentially" is growing exponentially. But only metaphorically. Literally, it's growing linearly.

3

38

191

Allen Downey

@AllenDowney

3 years

Without looking at the color bar, can you tell which colors are hotter than others? Now look at the color bar... does it help?

42

24

185

Allen Downey

@AllenDowney

4 years

This fall, I am taking my Data Science class on the road... ...the really long road to Ashesi University in Ghana: Classes start August 31. I can't wait!

9

17

184

Allen Downey

@AllenDowney

6 years

Every time someone invites me to Slack, I spend 15 minutes figuring out what email address to use, I connect once, and then never use it again. I'm on about 20 channels, can't log into any of them, and have no idea how or why anyone uses it. Is it terrible, or am I just old?

25

10

180

Allen Downey

@AllenDowney

3 years

New chapter in Elements of Data Science: using resampling to compute sampling distributions and confidence intervals (and understand what they mean).

1

23

176

Allen Downey

@AllenDowney

4 years

I ran into a NumPy gotcha today: np.var and np.cov have inconsistent default behavior. var divides by N cov divides by N-1 IMO, both should use N, which computes a simple descriptive statistic; if you want an estimator, you have to ask for it

numpy.cov() and numpy.var() default bias are inconsistent (numpy 1.9.2) · Issue #5835 · numpy/numpy

cov() uses a bias of 1 by default. var() uses a bias of 0 by default. Such that import numpy as np x = np.random.rand(100) if np.isclose(np.cov([x,x])[0,0], np.var(x)): print("Consistent by de...

github.com

9

28

175

Allen Downey

@AllenDowney

4 years

Theorem: If students find a statistical concept hard, the problem is the concept, not the students. Proof by example: p-values, confidence intervals and likelihood functions are "hard to understand" because they are fundamentally broken, bad ideas.

8

25

164

Allen Downey

@AllenDowney

5 years

Last week I asked for your favorite data visualization resources. Thank you to everyone who replied. I have organized the responses (and added a few of my own):

The library of data visualization

Getting ready for my Data Science class (starting next week!) I am updating my data visualization library, looking for resources to help students learn about visualization. Last week I asked Twitte…

www.allendowney.com

1

41

165

Allen Downey

@AllenDowney

4 years

What's new in English version 20.20? * Singular "they" is recommended. * Ending a sentence with a preposition is allowed. * Split infinitives are allowed. * "Whom" is now deprecated. * Latin plurals are deprecated. * The passive voice in science writing is deprecated.

7

35

159

Allen Downey

@AllenDowney

4 years

Here's a quick Bayesian analysis of the results from the vaccine trial announced today: Based on some guesses about the raw data, and some modeling assumptions, it seems unlikely that the effectiveness is less than 80%.

vaccine.ipynb

Run, share, and edit Python notebooks

colab.research.google.com

1

37

154

Allen Downey

@AllenDowney

1 month

I was at @Google today to give a talk about Chapter 7 of Probably Overthinking It: Causation, Collision, and Confusion. I'll post the video when it's available, but in the meantime, the slides are here:

1

14

156

Allen Downey

@AllenDowney

2 years

conda is so slow it is now unusable, and now can't solve some environments it used to. mamba is fast but buggy and the documentation is not ready for prime time. Is there a good option for package management?

51

10

155

Allen Downey

@AllenDowney

10 months

There are lots of good articles explaining how LLMs work at a mechanical level. This is the best explanation I've seen of how LLMs are able to do what they do, at least as we currently understand it.

Large language models, explained with a minimum of math and jargon

Want to really understand how large language models work? Here’s a gentle primer.

www.understandingai.org

3

33

149

Allen Downey

@AllenDowney

2 years

The Think Bayes red mullet is going to India! It looks a little nervous, but I'm sure it will be fine.

6

8

149

Allen Downey

@AllenDowney

8 months

3. But if you group people by decade of birth, most groups get more liberal as they get older.

4

28

149

Allen Downey

@AllenDowney

3 years

The second edition of Think Bayes went to the printer today! ebook versions will be available within a week, paper copies 2-3 weeks.

8

11

147

Allen Downey

@AllenDowney

3 years

I wrote an article that uses Bayesian decision analysis to find the optimal strategy for plugging in a USB connector. It turns out there's a reason it's so common to flip twice.

3

35

146

Allen Downey

@AllenDowney

5 years

My revised tutorial on Bayesian Statistics is ready to go. The slides and notebooks are here: If you are coming to @SciPyConf and you want to see it live, good seats are still available:

2

43

147

Allen Downey

@AllenDowney

4 months

I am excited to announce the forthcoming third edition of Think Python! What's new? Jupyter notebooks on Colab, learning to program with ChatGPT, regular expressions, automated testing -- and turtle graphics that work in notebooks!

10

18

147

Allen Downey

@AllenDowney

5 years

Does anyone know why GitHub has such a hard time rendering Jupyter notebooks? For me, it often fails several times and then works, or never works. Whereas nbviewer seems to work, quickly, 100% of the time. @github , can you borrow nbviewer's renderer?

12

17

143

Allen Downey

@AllenDowney

6 years

Reproducible Science Starter Kit: The Open Science Framework @LorenaABarba 's reproducibility manifesto Best Practices for Scientific Computing Software Carpentry; lessons learned

Best Practices for Scientific Computing

We describe a set of best practices for scientific software development, based on research and experience, that will improve scientists' productivity and the reliability of their software.

journals.plos.org

1

91

143

Allen Downey

@AllenDowney

4 years

If you've ever been confused about joint distributions, marginal distributions, and conditional distributions, you might like the next notebook in Bite Size Bayes: Welcome to the world of two dimensions!

10_joint.ipynb

Run, share, and edit Python notebooks

colab.research.google.com

2

26

141

Allen Downey

@AllenDowney

3 years

I'm teaching Complexity Science this semester, so I updated the notebooks. They run on @googlecolab , so you can run them without installing anything! Links here:

3

18

142

Allen Downey

@AllenDowney

1 year

I'm starting a new job this week, as a curriculum designer at Brilliant @brilliantorg , focusing on data science and computer science. If you want to try one of their online classes, here's a freebee:

14

4

140

Allen Downey

@AllenDowney

5 years

My thoughts on this whole 10x engineer thing: Many people are effectively 0x engineers (that's "zero ex") because they are working on things that will never have positive impact. Pay less attention to 10x; focus on making sure you are not 0x.

6

23

135

Allen Downey

@AllenDowney

1 year

The video of my tutorial on Bayesian Decision Analysis, from PyData Global 2022, is available now. For links to the video, slides, and Jupyter notebook, start at

2

36

137

Allen Downey

@AllenDowney

5 years

I am developing examples of survival analysis for my Data Science class. Anyone have any favorite application domains and/or datasets? @Cmrn_DP So far: 1) Literal survival probably using datasets from R 2) Time until marriage, divorce, NSFG 3) Customer conversion, dataset?

36

18

132

Allen Downey

@AllenDowney

6 years

I am moving toward treating NumPy as core Python and using all NumPy functions instead of the math module. The side effect is that Python sequences get quietly upgraded to NumPy arrays, which are usually faster and smaller.

Scientific Python

@SciPyTip

6 years

You cannot call math.cos() on a list, but you can call numpy.cos() on a list.

3

8

35

6

30

132

Allen Downey

@AllenDowney

4 years

I am working on an ebook, tentatively called Bite Size Bayes, that introduces Bayesian statistics gradually, for people with no prior stats. Here's Python notebook 2, if you want to check it out. R version coming soon!

02_bayes.ipynb

Run, share, and edit Python notebooks

colab.research.google.com

3

34

134

Allen Downey

@AllenDowney

2 years

Replication crisis update "none of the 193 experiments were described in sufficient detail" "of the 50 experiments from 23 papers that were [repeated], effect sizes were, on average, 85 percent lower than those reported in the original experiments."

Dozens of major cancer studies can't be replicated | Science News

A project aiming to reproduce nearly 200 top cancer experiments found only a quarter could be replicated.

www.sciencenews.org

3

46

130

Allen Downey

@AllenDowney

2 years

Really nice introduction to probablistic programming languages by showing a simple Python implementation.

A probabilistic programming language in 70 lines of Python

my blog

mrandri19.github.io

1

21

129

Allen Downey

@AllenDowney

4 years

Beryl Hoffman has written a Jupyter notebook that runs on Colab, pulls live COVID data, and demonstrates Pandas visualization tools. Really nice work!

Intro to Python Pandas with live COVID19 data.ipynb

Colaboratory notebook

colab.research.google.com

0

43

130

Allen Downey

@AllenDowney

5 years

A few weeks ago I led a workshop at Harvard on "Using computation to teach everything else" The slides for the workshop are at Including my favorite provocative slides:

4

40

126

Allen Downey

@AllenDowney

2 years

Just ran some older code and got a million warnings about features that have been deprecated. If you find yourself writing one of these warning messages, please, please, please include instructions for whatever the new thing is that replaces the deprecated thing.

9

11

121

Allen Downey

@AllenDowney

4 years

Mental health tip: DO NOT watch the election live like it's a sporting event. To keep you watching, TV people will make it seem like new information is streaming in. It's not. They're just making up stories about noise. If you must, check once Tuesday night. Then go to bed.

5

13

120

Allen Downey

@AllenDowney

5 years

Just used Tabula to read a data table from a PDF. Easy install; worked the first time. 10/10

5

10

117

Allen Downey

@AllenDowney

10 months

Let's practice Bayesian thinking. H: aliens D: unsupported testimony before Congress P(D|H) = high if there were aliens, someone would talk P(D|not H) = high not even the craziest thing said in Congress this week Likelihood ratio close to 1 = little or no evidence

Ed Krassenstein

@EdKrassen

10 months

Breaking: Former US Intelligence agent David Grusch, while testifying to Congress in the UFO hearings, just scared the crap out of me. The idea of UFOs and aliens isn’t frightening to me, but the idea that they could exist and that they could be trying to harm humanity is very

5K

2K

14K

4

10

112

Allen Downey

@AllenDowney

3 years

I just published an article that demonstrates my incremental process for developing and testing models in PyMC: It also explains the relationships between the four distributions of Bayesian analysis.

3

29

110

Allen Downey

@AllenDowney

5 years

In recent Jupyter , it looks like the default matplotlib backend is inline, so the magic command %matplotlib inline is no longer necessary. Good: that's one less thing to explain to beginners. But can someone confirm that I can count on this behavior? @ProjectJupyter

5

10

111

Allen Downey

@AllenDowney

6 years

I just learned about a new open-source statistics book by @russpoldrack It uses resampling/bootstrap for statistical inference, so I approve :)

0

26

112

Allen Downey

@AllenDowney

3 years

I think this example was written in Pascal, translated into Java, and then translated into Python. By someone who doesn't know Python.

10

12

109

Allen Downey

@AllenDowney

1 year

Here's a quick blog post summarizing my experiments: I stand by my conclusion: The great majority of coding will be LLM-assisted, starting now.

LLM-Assisted Programming

I’ve been experimenting with programming assisted by Large Language Models (LLMs) like ChatGPT. I am amazed at how good it is, and it seems clear to me that the great majority of programming …

www.allendowney.com

10

18

110

Allen Downey

@AllenDowney

4 years

I'm developing an introductory data science curriculum. Would you like to play along? Try out this do-it-yourself, choose-your-own-adventure mini-project that explores the relationship between political alignment and other attitudes and beliefs.

04_worldview.ipynb

Run, share, and edit Python notebooks

colab.research.google.com

3

25

109

Allen Downey

@AllenDowney

1 year

Exploring the effect of researcher choices on statistical results: 73 teams estimate the same effect size with the same data, and generate 1,253 different results. As always, statistical results depend on modeling decisions.

Observing many researchers using the same data and hypothesis reveals a hidden universe of uncert...

This study explores how researchers’ analytical choices affect the reliability of scientific findings. Most discussions of reliability problems in ...

www.pnas.org

5

33

110

Allen Downey

@AllenDowney

4 years

Here's a concise, readable summary of what we know about Covid-19, from the most reliable source we've got. Highly recommended:

2

43

107

Allen Downey

@AllenDowney

4 months

We are born Gaussian, but we grow up to be lognormal. Video, slides, and Jupyter notebooks from my talk @PyDataGlobal : Extremes, outliers, and GOATS: on life in a lognormal world

4

15

103

Allen Downey

@AllenDowney

3 years

New cover idea for Elements of Data Science. (The cover wraps, so the left half is the back and the right half is the front) What do we think?

13

10

105

Allen Downey

@AllenDowney

4 years

Just heard from my friends @OReillyMedia that Think Bayes second edition is a go. So, let's celebrate with the notebook for Chapter 2, featuring cookies, dice, socks, and Elvis.

Google Colab Notebook

Run, share, and edit Python notebooks

colab.research.google.com

2

12

104

Allen Downey

@AllenDowney

5 years

Last week I asked "If you were starting a book project today, what tools would you use?" I got many excellent suggestions, which I summarized here:

How to write a book

All of my books were written in LaTeX. For a long time I used emacs to compose, pdflatex to convert to PDF, Hevea to convert to HTML, and a hacked-up version of plasTeX to convert to DocBook, whic…

www.allendowney.com

4

30

102

Allen Downey

@AllenDowney

3 years

If you search for "python breadth first search", a substantial majority of the implementations you find are accidentally quadratic. Should be O(n+m), instead they are O(n^2). Here's the first hit: can you spot the performance error?

5

9

103

Allen Downey

@AllenDowney

7 years

More Python, less R, less Other.

6

69

95