Thomas Brambor

Slides from my #rstudioconf talk are up on my site! #0" target="_blank" rel="nofollow noopener">williamrchase.com/slides/assets/…

@arthur_spirling One cause of differences could be how different implementations of OLS deal with rank deficiency. If you have lots of categories etc. I could imagine this playing a role depending on how collinear things are dropped etc.

Pro-tip: nothing will reacquaint you with the foundations of the R and Python languages faster than having scikit-learn spit out *ever so slightly* different answers than lm() for the same data when you have a lecture to give tomorrow.

Lovely. An interactive version would be great.

I admire Kerry's ability to stay in that mind-bogglingly stupid conversation this long. Rep. Massie -- congratulations! You will from now on be my go-to example for dumb politician.

@Duke_of_neural The relatively short (7-weeks) insight data science program @InsightFellows has helped >2000 academics make the jump to industry. Perhaps an option to consider.

@jburnmurdoch Ah, thanks for this! Great memories of Raul, Klinsmann, Baggio, Del Piero etc. Would be nice to see Pele as a one off historical comparison. Though that may invite others to ask.... ;-)

Update to "top scorers in Champions League history": bit.ly/2HV3IGA You can now compare goals in *all competitions* since 1980 for ~ 200 players including • Salah, Kane, Aguero • Shearer, Baggio, Batistuta • Lineker, Klinsmann Also added player search, per-90 stats

@fhollenbach Looks interesting, but it seems to lack R Markdown support -- and I have to admit that is a killer feature in RStudio that I do not want to miss.

@mobydivingdick @albertocairo Actually, Africa is least distorted (since close to the equator). It's the countries you compare it to that are distorted by our Belgian friend Mercator.

@albertocairo in 2018, Mercator would be ridiculed as "Fake News" kind of guy. Below is the real size of Africa, the continent most distorted by his projection.

@Twiplomacy You guys must be loving this headline!

Great way to improve your regex game! ⚒️ twitter.com/grrrck/status/…

@arthur_spirling @kenbenoit @kmmunger Great. Nice to see the R package available right away! And just in time for teaching text mining in my grad courses.

Measuring *political* complexity w education reading index frm 1940s? STOP! New ppr + softwre @kenbenoit @kmmunger papers.ssrn.com/sol3/papers.cf…

And for those who care. Here is the R code to get the data on Brazilian elections since 1994 (0.5GB) and make the quick plot: gist.github.com/tbrambor/65fc2…

In honor of #WomensDay here is a quick #dataviz of the proportion of female voters in #Brazil over time. Looks like the 50 percent mark was just broken. Now it is time to get more #womeninpower

Great tutorial on dot density maps. Currently teaching a course on Dataviz - so this will linked in the course notes on maps. Thank you! twitter.com/tarakc02/statu…

@dataandpolitics Care to mention which school? I remember that Columbia students would mock NYU fans during sporting events for "ads in the subway". So, I want to have this down for my next niche joke at a polisci conference.

Want to learn a bit about the history of #Dataviz. Look no further. See the historical greats @ infowetrust.com/scroll/ twitter.com/infowetrust/st…

@maartenzam @lisacrost Nice idea @lisacrost. But this only accounts for part of the problem. Area sizes are still taken at face value, while the map projection further distorts our map visualization. It would even better to correct for the usual Mercator projection as well: thetruesize.com

Great #dataviz of the history of German street names. Informative, effective map visualization. Love it! twitter.com/venohr/status/…

The age of borders in the world. Really nice map - just kind of wish they did not use a rainbow color scheme. twitter.com/bill_easterly/…