2019
Vim anti-patterns, via @_wilfredh. #vim
Generative and Analytical Models for Data Analysis. Nice read!
For both the generative model and the analytical model of data analysis, the missing ingredient was a clear definition of what made a data analysis successful.
Also, be sure to check P values are just the tip of the iceberg (PDF).
Tidyverse pipes in Pandas. #python
🎥 Seven Sisters. Because I loved her role role as Lisbeth Salander (and because this one of the few films I’ve seen alone in the cinema).
I may have already posted this: Relearning Matrices as Linear Functions. If not, my bad, and here you are!
While I loved using nteract, and latter hydrogen on Atom, I’m back to basic Jupyter notebook when I have to. I noticed that the interact team recently released papermill, which can help parameterizing, executing, and analyzing Jupyter Notebooks. #python
The Tidynomicon, by Greg Wilson. Must-have ressource for those interested in switching to R, with a background in Python (or basically any scientific PL). #rstats
Since I’ve been using VS Code for a few days, this probably is a good read for tonight:
Foundations of Database. Didn’t know we had this from the INRIA team! (via @CompSciFact)
Long time Emacs user here: I’ve been using VS Code and NeoVim for five days. I’m perfectly fine.
Guess who tried to customize lintr for more than 30 minutes with syntastic while the plugin was globally disabled and I am using ale instead?
news-please – an integrated web crawler and information extractor for news that just works. #python
Just curious so I installed the datatable Python package. I’ve been very happy with R data.table package so I hope it will not be too difficult to switch over Python. BTW, the h2o.ai team is doing great things for data science. #python
Why Are Big Data Matrices Approximately Low Rank?. (via @ @GCLinderman)
I guess when you have some time to spare on a Sunday evening, you can just push a bunch of commits using the wrong user <email> values (probably updated by SourceTree that I just relaunched for the first time two days ago), right? In this case, this was very helpful. For fancier version, see How to change the commit author for one specific commit?
Visualising intersecting sets of twitter followers. #dataviz
Domain-specific languages are the ultimate abstractions. — Paul Hudak (1998)
hub is “an extension to command-line git that helps you do everyday GitHub tasks without ever leaving the terminal.” I have a working copy, of course, but I don’t use it often – StackOverflow to the rescue. Maybe it’s time to take a closer look…
📖 Iegor Gran, L’Écologie en bas de chez moi (P.O.L., 2011)
Always nice to read Zachary Tellman’s code on impure functional data structures, even if Java is not my business.
If you are looking to convert your Python notebook to a slide deck (and you like JS-based slideshow, of course), RISE looks like a good option. There’s also a PDF backend. #python
Looks like I just spend my full working day using VS code for Python coding again. It’s perfectly fine for building website (small codebase) or playing interactive script (< 200 LOC). Now, are there any better ways to integrate Jupyter notebooks than this? #python
Very nice showcase by the Stitch team.
With all DC teachers who have been out of business, we now have a plethora of nice tutorial comping up here and there. Here is one by Julia Silge on “tidy” text mining, and here is Ines Montani’s Advanced NLP course with spaCy. #rstats #python
A whirlwind intro to Python: A very nice intro to Python, written using Org. #python
I really like Yann Holtz’s teaching material, especially his Data analytics and visualization track. #dataviz
The Carpentries Handbook is live. Go check it if you’ve even been interested in teaching ressources.
I’m slowly, very slowly, updating the stata-sk project. on the one hand I am not in the best possible state of health, on the other hand my degree of motivation bought dramatically in recent weeks. I’m afraid it may have to end up in the garbage one day or the other.
At this point (order doesn’t matter, or I guess so), I wonder how this whole DC thing is going to end.
🎥 Highlander.
Folllowing the recent scandal around DataCamp, I think DataQuest and Exercism remain good online training centers for anyone eager to learn.
Composing Programs and From Python to Numpy are two of the most invaluable
ressources I know if you want to learn Python for real. #python
Great. Just when I was trying to recompile my Hugo website I noticed that the
0.54->0.55 upgrade just broke everything! And we can see comments like this. I
don’t care about performance issue or variable name changes or even Go itself. I
just want a tool to build my static pages, using an old theme that I customized
to my liking. Anyway, since I no time to fix the Go code for the modified
template I use, let’s go back to the previous release (0.54 has SHA
6c0c7919de42ee5d629d3a9786fb111f4498dab3) and pin it for good!
brew uninstall hugo
brew install https://bit.ly/2UxujU1
brew pin hugo
Nothing fancy on Twitter today. The web UI is just poor shit. Try to go back in your browser history, and a different series of posts appears on the very same page you landed on just a few seconds before. Oh, and did you ever manage to find an old posts of yours?