droodman / julia.ado

Bridge from Stata to Julia

julia.ado

This Stata package gives Stata users access to Julia. It contains three kinds of tools:

1. A jl: prefix command that lets you send single-line commands from the Stata prompt to Julia and see the results. Or, if typed by itself, jl starts an interactive Julia session in Stata.
2. Subcommands such as jl save and jl use for copying data between Julia and Stata.
3. Julia functions for reading and writing Stata variables, macros, matrices, and scalars.

Installation

ssc install julia

Documentation

After installing, type help jl. Also see the working paper in the /doc folder.

Requirements

• Julia 1.9.4 or later, installed following the instructions obtained via help jl in Stata after installing this pacakge.
• Stata 14.1 or later

Examples

. jl: "Hello world!"
Hello world!

. jl: sqrt(2)
1.4142135623730951

. jl: X = rand(3,100); X+X
3×100 Matrix{Float64}:
 0.708848  1.88261    0.600082  …  1.8036   0.660445  1.40321  1.98992
 1.21193   1.64774    0.389649     1.04665  0.584996  1.88493  1.50712
 0.701329  0.0138349  1.9605       1.35383  1.77841   1.93254  1.26002

. sysuse auto
(1978 automobile data)

. jl save auto   // copy data set to Julia DataFrame called "auto"

. jl: using GLM  // load generalized linear regression package

. jl: m = lm(@formula(price ~ mpg + headroom), auto)
StatsModels.TableRegressionModel{LinearModel{GLM.LmResp{Vector{Float64}}, GLM.DensePredChol{Float64, LinearAlgebra.CholeskyPivoted{Float64, Matrix{Float64}, Vec
> tor{Int64}}}}, Matrix{Float64}}

price ~ 1 + mpg + headroom

Coefficients:
─────────────────────────────────────────────────────────────────────────
                 Coef.  Std. Error      t  Pr(>|t|)  Lower 95%  Upper 95%
─────────────────────────────────────────────────────────────────────────
(Intercept)  12683.3     2074.5      6.11    <1e-07   8546.88   16819.7
mpg           -259.106     58.4248  -4.43    <1e-04   -375.602   -142.61
headroom      -334.021    399.55    -0.84    0.4060  -1130.7      462.658
─────────────────────────────────────────────────────────────────────────

. jl: SF_scal_save("adjR2", adjr2(m))

. display adjR2
.20542069