ky @kylefbutts Twitter profile

Pinned Tweet

ky

2 years

Mixtape Sessions 🥳FREE🥳 library now has 4 complete courses (and our revamped diff-in-diff/synthetic control course is a WIP) --- What we have ...

1

78

331

Last Seen Profiles

@seeksmyperson

@IsraelIstanbul

@SundayTimesZA

@kmk20062013

@ScrollaAfrica

@jnsoofolder

@fabrice_folio

@TerrivaMacAlist

@isucyclone

@by4mao

@HedgeComber

@ren_shibasui507

@_Palomolean

@Nust

@AlinaKrylovaaa

@Bn_gsh2

@wbeimarlodice

@JuanJua04542850

@AlexanderTaffy

@thegirl_russia

@SenaMara16697

@Alkhdashmohamm1

@JaidTaylorLive

@GodrejEntGroup

@hope_MKII

@Mauisnake

@Shingo_Raiu

@smirusnews

@Eduartdirection

@boboobul

@SchiefAili

@BitcoinOllie

@___doxxy

@SimonDixonTwitt

@minjoounion2024

@TulioBraga999

ky

@kylefbutts

3 years

R has a reputation of being terrible at running regressions. This is no longer true, thanks to -fixest- package. I made a Stata -> R Cheatsheet to try and prove it @lrberge single-handedly remade all the good features of stata. It's truly a feat

34

301

2K

ky

@kylefbutts

3 years

🎁New gift alert:🎁 The entire diff-in-diff robustness check in one R function: Beta mode, so please report any bugs if you try it!!

7

140

812

ky

@kylefbutts

3 years

Introducing w @nickchk and @grant_mcdermott Hopefully makes the Stata -> R pipeline much easier. - data-wrangling with {data.table} - regression analysis with {fixest} These packages are blazingly fast 🏃🏻💨, highly powerful 💪🏻😎, and flexible 👌✨

12

150

742

ky

@kylefbutts

3 years

🚨 Beamer Slides Template 🚨 My attempt at making beautiful beamer slides easy. h/t to @paulgp 's beamer tips for a lot of inspiration. pdf preview: source code: Details...

8

100

741

ky

@kylefbutts

3 months

Michal Kolesar's Advanced Applied Econometrics lecture notes are a delight

4

159

745

ky

@kylefbutts

1 year

Semestrial reminder that *all* our workshop materials at is available for *free* on GitHub (and growing):

4

151

726

ky

@kylefbutts

3 years

How I go from historical tables on the left -> pretty clean csv data on the right?

9

66

461

ky

@kylefbutts

2 years

Built this for @causalinf 's Mixtape Sessions next week. As always all this work is generating open source material. Build some intuition on optimal bandwidth selection @maxhfarrell

6

62

457

ky

@kylefbutts

3 years

#EconTwitter , I don't know if folks know this, but Davidson and MacKinnon's two books are free to download online now Econometric Theory and Methods: Estimation and Inference in Econometrics:

7

140

431

ky

@kylefbutts

9 months

While prepping my JMP slides, I've made some pretty cool updates to my latex slides template: - color rows/cells (with reveal options) - bottom left beamer buttons: \bottomleft{} - customize theme colors(!) - full-size figures: \imageframe{}

9

42

401

ky

@kylefbutts

2 years

@causalinf and I finished the first half of our flagship causal inference course. All the material (besides Scott of course) is *available online* and *free* for anyone to use 🥳🥳

GitHub - Mixtape-Sessions/Causal-Inference-1: Causal Inference 1 Mixtape Session taught by Scott...

Causal Inference 1 Mixtape Session taught by Scott Cunningham - Mixtape-Sessions/Causal-Inference-1

github.com

2

91

397

ky

@kylefbutts

4 years

@causalinf and I compiled all the R scripts from his new book and created teaching assignments that can be run directly from R Studio with questions for students. Here is an example from his chapter on Regression Discontinuity:

4

72

376

ky

@kylefbutts

2 years

🌠🔭I wrote a blog post about all the weird diff-in-diff settings people email and ask me about (continuous treatment, treated multiple times, treatment turning on and off)

3

54

357

ky

@kylefbutts

2 years

So happy with how Brigham Frandsen’s Machine Learning for Causal Inference workshop turned out. All the material, per usual, is free to use!!

GitHub - Mixtape-Sessions/Machine-Learning: Machine Learning and Causal Inference taught by Brigham...

Machine Learning and Causal Inference taught by Brigham Frandsen - Mixtape-Sessions/Machine-Learning

github.com

4

79

346

ky

@kylefbutts

3 years

New WP 🎷🎷 The ring method: when treatment occurs in a specific location, use inner-ring as treated & outer-ring as control group. Researchers use this all the time, but no one has formalized the necessary assumptions... *horns intensify* 👉

12

57

344

ky

@kylefbutts

6 months

This fall, I have taken a job as an AP at the University of Arkansas !! As I have done (annoyingly) around the house for the last week: WOOOO PIGGGG I'm so thankful to all those that have helped along the way:

39

3

295

ky

@kylefbutts

3 years

I just posted my first working paper, Difference-in-Differences Estimation with Spatial Spillovers. Another diff-in-diff paper? 🙃🙃 Let me explain...

Difference-in-Differences Estimation with Spatial Spillovers

Empirical work often uses treatment assigned following geographic boundaries. When the effects of treatment cross over borders, classical difference-in-differences estimation produces biased...

arxiv.org

7

58

293

ky

@kylefbutts

3 years

New WP🤌 Often, people use Regression Discontinuity with some administrative border as the cutoff. This can be bad if: 1. People sort on either side of the border (better schools or lower taxes) 2. Many laws change at border...

7

54

280

ky

@kylefbutts

10 months

Really great timing on this R&R 🥳

10

16

271

ky

@kylefbutts

3 years

🚨Latex Article Template🚨 Folks really seemed to like my slides, so here's the article template I have been using. pdf preview: source code:

4

40

267

ky

@kylefbutts

3 years

PSA:

9

19

261

ky

@kylefbutts

3 years

Opinionated Latex table tip: You should output only the body of the table in R/Stata and do everything else in latex. This makes it simple to set up multicols, alignment, footnotes, etc. without trying to learn all the stata/R options.

10

32

255

ky

@kylefbutts

3 years

Want to see visually the problems with TWFE Diff-in-diff with differential timing? Try out a shiny app and see when TWFE works particularly poorly: If you use RStudio, save compute time for other ppl and run locally plz:

3

52

252

ky

@kylefbutts

2 years

🚀 What do you think the best written econ paper you've ever read is? 🚀 PLEASE submit it here. Mixtape Sessions is trying to make a catalog of > 100 awesome papers folks can look through to inspire their writing

11

36

233

ky

@kylefbutts

3 years

#EconTwitter If you have not used pandoc before, it's really remarkable. In one line, I went from latex to docx and it works amazingly. Even matches the formatting of my custom latex template pandoc paper.tex -o paper.docx

6

34

227

ky

@kylefbutts

10 months

stata2R, our guide on learning R for stata developers has fresh new paint powered by @astrodotbuild : The built in search is fantastic for quickly finding a topic @grant_mcdermott @nickchk

4

50

207

ky

@kylefbutts

3 years

WIP: Event-study estimates in R all in two commands. Just need to add did_imputation by @borusyak et. al.

6

16

179

ky

@kylefbutts

3 years

Just got my first R&R !! On a Friday night, no less

ky

@kylefbutts

3 years

New WP🤌 Often, people use Regression Discontinuity with some administrative border as the cutoff. This can be bad if: 1. People sort on either side of the border (better schools or lower taxes) 2. Many laws change at border...

7

54

280

7

4

174

ky

@kylefbutts

2 years

🎉 New Working Paper: This paper, joint with Nick Brown, shows how to estimate treatment effects *when* parallel trends fails Full paper & 5 Minute Summary:

3

35

174

ky

@kylefbutts

3 years

new tools from the metrics world 🔨🔨 I'm a big fan of John Gardner's new working paper on two-stage diff-in-diff and made it into an R package. It fixes TWFE estimators for static and event study estimates! install here:

GitHub - kylebutts/did2s: Two-stage Difference-in-Differences package following Gardner (2021)

Two-stage Difference-in-Differences package following Gardner (2021) - kylebutts/did2s

github.com

4

31

168

ky

@kylefbutts

2 years

This is really exciting. You can read a 5-minute summary of this paper here (on the new version of my website):

Kyle Butts

Kyle Butts -- Urban Economics and Applied Econometrics

www.kylebutts.com

Journal of Urban Economics

@JUrbanEcon

2 years

New in #JUE : "JUE Insight: Difference-in-Differences with Geocoded Microdata" #UrbanEcon #EconTwitter

2

36

162

8

27

153

ky

@kylefbutts

3 months

With another course coming to a close, all the course material is now a public good!!! I might summarize @instrumenthull 's class as "we think something random is occurring, how do we use this to estimate effects?"

GitHub - Mixtape-Sessions/Design-Based-Inference: Design-Based Inference Mixtape Session taught by...

Design-Based Inference Mixtape Session taught by Peter Hull - Mixtape-Sessions/Design-Based-Inference

github.com

1

31

152

ky

@kylefbutts

3 years

Would #rstats folks read/watch me working on Rcpp programs that speed up my workflow? I think there's perhaps a missing supply of beginner Rcpp usage

13

5

148

ky

@kylefbutts

1 year

Wrote a blog post about how understanding factor models help me understand the synthetic control method much better: 👉🏼

5

29

144

ky

@kylefbutts

3 years

🥳 New WP with John Gardner introducing the R package `did2s` to estimate TWFE/Event-study with staggered treatment article: If anything is confusing please let me know :-) 1/n

1

25

140

ky

@kylefbutts

2 years

New blog post showing why with `did2s` you don't need to include individual fixed effects (instead just treatment-timing group FEs). You'll get numerically identical results from @jmwooldridge 's new paper! Major speedups for people with big data!!

Kyle Butts

Kyle Butts -- Urban Economics and Applied Econometrics

www.kylebutts.com

2

30

131

ky

@kylefbutts

2 years

Working on interactive Shiny apps for @causalinf 's Causal Inference mixtape session next week. Check out this WIP for randomization inference #RStats #Shiny

1

13

131

ky

@kylefbutts

2 years

Announcing a new online seminar series for graduate students doing urban (or urban adjacent) economics organized by @santiagohermo , @_Hector_Blanco , @Jon_Hartley_ , @laurayww and myself Sign-up to attend or to present!!

RUSH (Regional / Urban / Spatial / Housing) Graduate Student Brownbag

grad-student-urban-brownbag.netlify.app

2

51

129

ky

@kylefbutts

8 months

WIP: Write R scripts like you normally do and have them automagically turn into a git-friendly log book that you can share with your coauthors using @quarto_pub

1

10

116

ky

@kylefbutts

7 months

log(x+1) is down. well well well, hello log(x+2) my new friend… Just kidding; this paper is a delight of a read. Crisp and clear

QJE

@QJEHarvard

7 months

Recently accepted by #QJE , “Logs with Zeros? Some Problems and Solutions,” by Chen ( @jiafengkevinc ) and Roth ( @jondr44 ):

4

195

726

2

5

110

ky

@kylefbutts

5 months

Guido Imben's review on using covariates for treatment effect estimation is great: Interestingly, the title changed from "Semiparametric" to "Nonparametric" Estimation when the published version at RESTAT came out 🤔.

0

28

102

ky

@kylefbutts

10 months

New package `fwlplot` is ready to be tried out by some volunteers! This is a 1:1 drop-in replacement for `feols` but instead produces a scatter plot of residualized y on residualized x (first x-var in formula). Works with fixest multi features, e.g.

1

13

99

ky

@kylefbutts

3 years

See problems with TWFE event-study estimates. Example:

2

11

99

ky

@kylefbutts

3 years

Easy standard errors, first-class post-estimation commands, lag/leads/differences, IV All this while being the fastest regression software out there!!

3

11

97

ky

@kylefbutts

2 years

Really cool paper in JUE this week. Say a nice new apartment opens up in your city, what happens? I'll tell you two plausible stories, and this paper (along with a similar paper by @EvanMast2 ) can point us to which seems to be happening

4

31

96

ky

@kylefbutts

2 years

I cracked a metrics problem that I've thought about for over a year and a half today. Feeling pretty stoked!!

2

0

94

ky

@kylefbutts

3 years

My function `did2s` is like 90% just a wrapper around @lrberge 's fixest package (w/ standard error adjustment). It is crazy good. It can do FEs faster than anywhere else, IV, GLM, export results to figures and tables incredibly easily. Check it out!!

2

13

90

ky

@kylefbutts

9 months

This course is a really great one. Developed this app to build intuition on linear shrinkage estimators (on simulated school value-added data) Shiny app: More info on the course:

Christopher Walters

@c_r_walt

9 months

Next week I'm teaching a Mixtape Session on Empirical Bayes. Register at the link below if you want to learn about using EB methods to quantify heterogeneity, "borrow strength" to improve treatment effect estimates, and make decisions:

1

15

128

1

13

81

ky

@kylefbutts

1 year

in our lovers era (btw we got engaged !!!)

5

0

80

ky

@kylefbutts

2 years

I’m looking for diff-in-diff papers where the author proposes a mechanism for how D affects Y (preferably with publicly available data)

5

12

76

ky

@kylefbutts

1 year

@paulgp @grant_mcdermott Update to R 4.3.0, and you can use `_[]` in pipes. I'm super pleased

7

1

73

ky

@kylefbutts

3 years

Code to implement two-stage diff-in-diff are here! Stata: R:

GitHub - kylebutts/did2s_stata: Two-Stage Difference-in-Differences following Gardner (2021)

Two-Stage Difference-in-Differences following Gardner (2021) - kylebutts/did2s_stata

github.com

scott cunningham

@causalinf

3 years

Finally finished my substack on John Gardner’s “two stage difference-in-differences”. I’ll do a brief summary here.

4

39

169

0

16

75

ky

@kylefbutts

8 years

Thanks for sharing sunshine and political revolution with us today :) @shailenewoodley @kendrick38 @rosariodawson

0

25

64

ky

@kylefbutts

2 years

If you want to update to Stata 17 without paying, use this trick: right click Stata and select "get info". Then drag and drop the Stata 17 icon. Taadaa ✨✨

0

2

68

ky

@kylefbutts

2 years

I'm really excited about this one. This Saturday, the first "frontier" course for Mixtape Sessions for Shift-Share Instruments. The material is beautiful (💅) and features a coding lab (⌨️) to get your hands dirty

Shift-Share IV - May 21st

Shift-Share IV. May 21st. Full Schedule: https://www.mixtapesessions.io/session/shift_share_may21

www.eventbrite.com

1

7

68

ky

@kylefbutts

1 year

Workers who reside in urban areas earn a large premium over non-urban workers. Is this driven by sorting or due to agglomeration effects of cities? And how does this change as we go back in time? New working paper w/ @kitchct and Taylor Jaworski: 1/N

2

20

64

ky

@kylefbutts

3 months

fixest tip: you can "attach" a custom variance-covariance object to a fixest estimate using `summary`. The custom vcov gets used with `vcov(x)`, so it works with `modelsummary`, `etable`, `iplot`, etc. e.g. with bootstrap standard error:

0

11

62

ky

@kylefbutts

3 years

Shiny app to see the @agoodmanbacon decomposition theorem: Two recommendations on things to try to gain insights into the paper ...

Pedro H. C. Sant'Anna

@pedrohcgs

3 years

So good to see another recent and influential DiD paper finally published in one of my favorite journals: Congrats @agoodmanbacon !!

4

51

390

1

16

59

ky

@kylefbutts

2 years

#RStats Started working on a mega blog post all about TailwindCSS and why I think it will make designing apps with #Shiny so much easier #WIP

2

7

58

ky

@kylefbutts

2 years

This might be the best course material we've made so far. Look at how good everything looks. Code with interpretation built-in and beautiful slides (if I say so myself)

scott cunningham

@causalinf

2 years

Four more days (Mar 21) until @instrumenthull workshop at Mixtape Sessions on shift-share ("Bartik IV"). If you are interested in topics even remotely within rocks throw of labor, immigration, trade and macro, highly recommend.

2

13

71

1

7

58

ky

@kylefbutts

1 year

Boston making monocentric city models proud 🥲

3

1

55

ky

@kylefbutts

1 year

We'll be talking about how to go from (a) to (b) next week :-) ...no guarantees it will work as well in your setting ;)

RUSH Brownbag

@RUSHBrownbag

1 year

Next Wednesday, April 5th at 11am EST, we're back! Kyle Butts present his work on imputation estimators in factor-models Sign-up to attend: Full Schedule:

0

1

15

1

4

58

ky

@kylefbutts

3 years

`did2s` has come to Stata now too. All the details here:

GitHub - kylebutts/did2s_stata: Two-Stage Difference-in-Differences following Gardner (2021)

Two-Stage Difference-in-Differences following Gardner (2021) - kylebutts/did2s_stata

github.com

ky

@kylefbutts

3 years

new tools from the metrics world 🔨🔨 I'm a big fan of John Gardner's new working paper on two-stage diff-in-diff and made it into an R package. It fixes TWFE estimators for static and event study estimates! install here:

4

31

168

1

13

55

ky

@kylefbutts

10 months

This is such a fun profession. This tweet led me to my JMP...

ky

@kylefbutts

3 years

#EconTwitter Do fixed effects still control for a time-invariant variable if the coefficient on the variable changes over time? E.g., wage on education. Work ethic doesn't change over time, but the returns to work ethic do.

3

1

14

1

54

ky

@kylefbutts

6 months

Been formalizing a lot of my workflow into R packages. Today, I have added `tikzsave` to `kfbmisc` package for making figures: ggplot() -> tikzpicture -> pre-compiled into pdf (using latex themes)

0

4

56

ky

@kylefbutts

2 years

Three-stage did prototype is functional now (the third stage is for time-varying covariates !!!) Does anyone want to help me debug and try with their project? DM me!

2

7

54

ky

@kylefbutts

3 years

#rstats is a game-changer for teaching R in blog posts. Just by adding comments, torchlight can highlight lines, make collapsible code, git-style diffs, focus lines (unblur on hover), and more. All with your VSCode theme!! Try it:

1

10

53

ky

@kylefbutts

3 years

#RStats Here's my submission to the RStudio Table Competition. Designed with @tailwindcss using my package shiny.tailwind. Tailwind makes it so easy to design shiny apps

2

12

52

ky

@kylefbutts

1 year

Happy (incredibly annoyed) to announce the 1-line update to did2s has hugely sped-up did2s for large datasets (something like 20x) 🤦🏼‍♂️:

0

2

53

ky

@kylefbutts

5 months

One step closer to dream of `NBER` ggplot2 theme :-)

Thomas Lin Pedersen

@thomasp85

5 months

One of the big improvements in the new version of #ggplot2 are related to legends. @TeunvandenBrand lays out all the detail in this blog post. Get ready for the legend extension extravaganza #rstats

0

33

173

1

6

49

ky

@kylefbutts

2 years

most data science is data cleaning. Over 1000 lines of code cleaning addresses 🫠

3

49

ky

@kylefbutts

3 years

#EconTwitter I'm looking for example papers that use this identification strategy: Treated units are in a treated boundary (or within a ring from treatment location) and Control units are just outside of the boundary (just outside this ring)

21

11

47

ky

@kylefbutts

3 years

I really love this paper, but I wanted to highlight one thing that I think @pedrohcgs (among others) gets totally right in this literature: Instead of starting with an estimator and try and figure out what you estimate, you should just start with an estimand! 1/2

Pedro H. C. Sant'Anna

@pedrohcgs

3 years

🚨Updated Working Paper Alert🚨 "Efficient Estimation for Staggered Rollout Designs", joint with @jondr44 (). This paper is all about how we can do better than DiD in setups with quasi-random treatment timing! 1/n

4

62

309

1

4

47

ky

@kylefbutts

5 months

@tcarpenter216 I help run . We have all our course material online on . We have a live course on diff-in-diff next weekend :-)

Mixtape-Sessions

Mixtape-Sessions has 21 repositories available. Follow their code on GitHub.

github.com

2

6

46

ky

@kylefbutts

11 months

ICYMI: Urban/Housing PPL 📣 The FHFA has a new tract-level repeat-sales housing price index (1989-2021) with >65,000 tracts using data from Fannie Mae and Freddie Mac, CoreLogic county recorder files, and FHA mortgages

1

17

45

ky

@kylefbutts

3 years

1. Using LayoutParser by @MelissaLDell and coauthors, I extract the actual table from the sheet

1

45

ky

@kylefbutts

10 months

I'm on the job market and can't let myself get distracted by dumb open source ideas, but it would be really cool to have an easy way in ggplot to put the legend like this and the y-axis label like this. NBER graphs always are 👨🏼‍🍳🤌🏼

3

44

ky

@kylefbutts

2 years

This Friday and Saturday, Mixtape Sessions is hosting Doing Applied Research. There's so many great insights in these slides! Perfect for junior researchers Full Schedule:

0

7

45

ky

@kylefbutts

6 years

I'M SO EXCITED TO READ @curaffairs

0

2

41

ky

@kylefbutts

3 years

Just wrote code in C++ in parallel using RcppParallel and code ran from 45 minutes to 3 seconds 🤯

2

0

42

ky

@kylefbutts

3 years

Can Econ just require ": Evidence from earth" if the author doesn't include a "Evidence from" statement in the title

2

0

41

ky

@kylefbutts

2 years

Helping @instrumenthull prepare for his IV workshop on Friday and he has a set of simulations for weak and many instruments that blew my mind 🤯

3

2

39

ky

@kylefbutts

2 years

@paulgp here is a little trick I learned:

1

2

40

ky

@kylefbutts

5 months

Night 1 of Demand Estimation with @jeff_gortmaker went fantastic. Ariel Pakes started us off with a history lesson on Demand Estimation and what came before BLP

0

4

38

ky

@kylefbutts

3 years

Haven't even put it out yet, but I just referenced my WIP data.table guide w @nickchk and @grant_mcdermott to remember how to drop duplicate observations

1

0

38

ky

@kylefbutts

2 years

The { here } package is awesome in R. Notice how weird of a location my script file is. here::here() gives me the absolute path relative to the root folder no matter where the script file is. A bit more details ⬇

AEA Data Editor

@AeaData

2 years

Please stop using "cd" (in Stata) or "setwd()" (in R) all over the place. Once (maybe, not really), that's enough. Can we mark those commands as "deprecated"?

21

63

636

2

3

38

ky

@kylefbutts

1 year

ChatGPT'd my way from an idea to a package in 1 hour: 🏎️💨 fast Rademacher weights in Rcpp

5

1

38

ky

@kylefbutts

2 years

@eckhoffandresen

1

2

36

ky

@kylefbutts

3 years

Used blastula and #Rmarkdown today to send out customized midterm reports for students. So simple, it took less than an hour from drafting to emailing. Reach out if you want to know how!

4

2

34

ky

@kylefbutts

2 years

Citing software you use is really really a good way to encourage people to code estimators for you ! In R, type citation(“package”) and it will tell you how to properly cite it!

Will Wheeler

@willwheels

2 years

@wytham88 That's a good point. I'd also think about the value of having public code vs. a formal package. A lot more work is involved w/ the latter, so it's great when people can do it (thanks @grant_mcdermott and @kylefbutts !!!). Adding Sun and Abraham to fixest is a great idea, e.g.

1

0

1

0

3

33

ky

@kylefbutts

3 years

Visual Studio Code is great. Here's some reasons:

scott cunningham

@causalinf

3 years

What text editor for programming are you guys using? I’m switching from Textmate 2 to one called Visual Studio Code bc @kylefbutts was raving about it a few months ago. I was using Textmate 2 for a while and loved it so may go back. Just curious.

25

1

35

3

4

34

ky

@kylefbutts

3 years

Hi #rstats , I'm teaching R this fall and was wondering if folks have material to recommend and/or experiences to share! :-)

8

6

33

ky

@kylefbutts

3 years

If you have panel/repeated cross-sectional data before and after the treatment, then you can use a Difference-in-Discontinuities estimator to fix these problems Yes Diff-in-Disc, not Diff-in-Diff :-)

1

2

32

ky

@kylefbutts

2 years

Besides @jondr44 , I might be the only researcher who's excited to have these bad of pre-trends

2

1

32

ky

@kylefbutts

2 years

Hell yeah, GitHub lets you use latex equations in markdown documents now:

Render mathematical expressions in Markdown

github.blog

0

5

32

ky

@kylefbutts

5 months

#EconTwitter If I run a regression with outcome y, a variable of interest x and controls W and I want to plot the variation, I use FWL and scatterplot y_resid on x_resid. What would you do in the case of IV?

6

4

32

ky

@kylefbutts

1 year

Officially Dr. @hannah_denker !! She’s the most intellectually-honest and caring researcher I know 💜

1

0

31

ky

@kylefbutts

2 years

I think the most fun part of the job is getting to try and explain fancy econometric methods in really intuitive ways. It feels like a big puzzle. #EconTwitter #WIP This is the paper I'm applying:

1

4

30

ky

@kylefbutts

2 years

After more than one year, I finally have closed this issue 🙏🏼🙏🏼

1

0

30