Twitter | Search | |
Peter Ellis
Statistician / data scientist with wide-ranging interests. Tweets represent my views (if that), definitely no-one else. Day job is in the public sector.
2,418
Tweets
1,451
Following
1,392
Followers
Tweets
Peter Ellis retweeted
Liz MacPherson May 16
Thanks !! pleased to be selected along with and . Looking forward to seeing a growing list of shiny apps and to see them being used!!
Reply Retweet Like
Peter Ellis May 15
Replying to @EmmaVitz
Congratulations!
Reply Retweet Like
Peter Ellis May 14
Replying to @emilynz
Yes
Reply Retweet Like
Peter Ellis retweeted
Stencila May 13
Great article! That's what we're aiming to provide with - a familiar interface (such as Stencila Sheets) empowered by reproducible analysis in R, Python, SQL, JavaScript - you name it. Would love to see you as our !
Reply Retweet Like
Peter Ellis retweeted
Maarten van Smeden May 13
In the last couple of years I’ve met many PhD student and early career researchers in doubt or discouraged to learn a statistical programming language (R, Python, Julia, …). I’ve made a list with 10 reasons to start programming. [THREAD]
Reply Retweet Like
Peter Ellis retweeted
Noah Smith May 13
New idea: Deal with a productivity slowdown by giving old things exciting new names! Statistics = "data science" Chemistry = "nanotech" Marketing = "growth hacking" Conservatives discussing conservative ideas = "Intellectual Dark Web"
Reply Retweet Like
Peter Ellis May 12
All the ones in that post are on 's , so they should scale up ok as necessary. Some agencies also use Shiny Server Pro for internal use but I don't know in NZ govt of it being used yet for public dissemination. Licensing model not quite right.
Reply Retweet Like
Peter Ellis May 12
A fistful of Shiny apps by New Zealand govt agencies for described at . Includes examples from , , ,
Reply Retweet Like
Peter Ellis retweeted
bletchley punk May 9
On self-guided programmers and CS grads
Reply Retweet Like
Peter Ellis retweeted
John Burn-Murdoch May 9
😍😍😍 this post on dual-axis line charts (discovered via That Awesome Article™). Most nuanced & detailed take I've seen on this topic, w/ excellent guidelines both on *whether* to use them, *and* best practice if you do
Reply Retweet Like
Peter Ellis May 7
Replying to @emilynz
Reply Retweet Like
Peter Ellis May 6
Yes, there may be new interest from the private sector in robust anonymisation and statistical disclosure control, as practiced by national statistical offices.
Reply Retweet Like
Peter Ellis May 4
Musings on good books, courses and tools from my last few months reading and thinking about them. Also, everyone should think in terms of "total survey error", regardless of whether their data is from a survey or not.
Reply Retweet Like
Peter Ellis May 3
Multi-indexing with pandas DataFrames in makes me feel like I am using Excel. Is that the desired effect?
Reply Retweet Like
Peter Ellis May 1
Couldn't that help snoopers trying to re-identify? I would have thought a random list was the only safe approach.
Reply Retweet Like
Peter Ellis retweeted
Shaida Badiee Apr 26
Wonderful news: a female Chief Economist for the World Bank. She will be the second one ever after Anne Krueger 82-86.
Reply Retweet Like
Peter Ellis Apr 19
or Malcolm Turnbull...
Reply Retweet Like
Peter Ellis Apr 19
Replying to @Economissive
Did you pass, or have you been found out to be a robot?
Reply Retweet Like
Peter Ellis Apr 19
Replying to @ellis2013nz
And the slides are available at
Reply Retweet Like
Peter Ellis Apr 19
Video is now available of my talk for the Wellington R users group, shortly before the last election on forecasting it from polling data with and .
Reply Retweet Like