Table of Contents

Announcement

I just discovered the low-end box that I used to automatically download my timeline has run out-of-memory, and all download tasks in the last 24+ hour have failed. This is an opportunity for me to hit the pause button.

This project started to help me keep track of interesting tweets that I might want to go back to later. To some extent, it has served that purpose well. However, even with automatic clustering and sorting, reading the tweets was still consuming more time than I’d like. And I’d been struggling to find an easy and robust way to present the tweets in the web pages. All in all, I need to take some time off to find out a better approach to this project. The potential solutions include:

  1. Use a whitelist of Twitter users in addition to universal encoding and clustering to further reduce the number of tweets to read.
  2. Develop a better UI and workflow for the tweet reading-marking-categorizing process.
  3. Explore other ways of presenting collected tweets other than grouping by date of appearance on my timeline.
  4. Reduce the scope. Maybe it’s better and easier to focus on one of the following topics: research, tutorial, ideas/thoughts, or visualizations.

Anyway, I still collected five tweets from my timeline and put them in the following section. Maybe they’ll be the last on this site, or maybe not. We’ll see. :)

Miscellaneous

I try to let my data speak for itself.

But it mostly just says, despondently, gazing off into the distance:

“Gelman would do such a better job with this analysis. <sigh> But I’m stuck with you, aren’t I?”#DataScience #statistics

— Nihilist Data Scientist (@nihilist_ds) July 18, 2018

Unpacking the polarization of workplace skills

Network science shows that skills exhibit polarization into two clusters that highlight the specific social-cognitive skills and sensory-physical skills of high- and low-wage occupations, respectivelhttps://t.co/wC3uvrecy9 pic.twitter.com/ZnNsKAi6Eh

— Alessandro Vespignani (@alexvespi) July 18, 2018

💖 @krstoffr's 💫 interactives!
"The t-distribution & its normal approximation: visualization" https://t.co/7T359n2nGN #infovis #statistics pic.twitter.com/tdo9bhlsUc

— Mara Averick (@dataandme) July 19, 2018

How Decision Trees Work is livehttps://t.co/xuQqiWfiR4 pic.twitter.com/dyVIR7BOJb

— Brandon Rohrer (@_brohrer_) July 19, 2018

Sentiment analysis of tweets from members of Congress reveals...
- Tweets from women tend to be more positive than tweets from men
- Sentiment of tweets by party has become noticeably more polarized since Nov 2016 pic.twitter.com/T2A0eksh41

— Mike Kearney📊 (@kearneymw) July 18, 2018

@ceshine_en

Inpired by @WTFJHT