#RStats #fixest

Dear `fixest` users: in v 0.10.3 I improved the demeaning algorithm (ye!)

It should be faster for complex cases (many FEs with difficult convergence)

But I only checked with my test cases.
Could a few good souls check it's faster for them too?

lrberge.github.io/fixest/refer...

Notwithstanding, I love that information on na-s in data. Or when #fixest tells you stuff like
235/94/0 fixed-effects (917 observations) removed because of only 0 outcomes.

(live tweeting research calls while code runs...)
So when have fixed effects (incl industry or country dummies), @Gabors_Data says there shall be no intercept/constant in regression table... Let's see:
#Rstats #Fixest: is the default 👏
@Stata #reghdfe: need ",noconstant" 👌

#Rstats tip. Suppose you have 5 countries in the dataset and wanna check how estimated regression coefficients vary. So easy in #fixest. `,fsplit = ~country` and same regression will be run on full sample and one by one on country subsets. @lrberge...

#Rstats #fixest etable question... So I want basically
*a table by etable as is (not transformed by kbl)
*with proper footnotes like in threepart table (which kbl can do, but I can't do with etable)
*saved as "folder/table1.tex"
Can I have it all? THX! @grant_mcdermott @lrberge

My new "hobby" is to bug authors of my favourite #Rstat packages with ideas on how they can spend their Fridays .(Github issues make it too easy.) Thanks, beautiful people. Like #fixest @lrberge and #modelsummary @VincentAB. Btw two super pkg for econ. #EconTwitter