Rln (ARLEN): free, offline statistics and data exploration for researchers, with two desktop editions and a native Android app.

Rln (ARLEN)

Free, offline statistics and data exploration for researchers — two flavors of desktop and Android.

v1.2.8 MIT licensed 100% offline Windows · macOS · Linux · Android
Download latest release View on GitHub

Android APK · Windows / macOS / Linux executables · or run from source

Pick your edition

Two flavors, one command language

The same engine and the same scripts run everywhere. Choose an edition to tailor the install steps and capability map below.

What you get

Full statistics & econometrics (regression, panel, IV, every diff-in-diff estimator), charts, larger-than-RAM streaming, fuzzy merge, and the color-coded data explorer. ~80–120 MB. No NLP (use the offline/full tiers for that).

Best for

Day-to-day analysis on a laptop or workstation — the build most people want. A graphical app and a command-line REPL in one download.

What you get

The full command language, native and offline, with a built-in NumPy/SciPy estimation engine: regress, logit, probit, poisson, panel fixed effects and DiD (TWFE). LRTM streams Parquet with no pyarrow, and a tap shows a live preview. Dark mode and an always-visible script editor.

Best for

Fieldwork and analysis on a phone — millions of rows summarized in a fraction of a second, with your data never leaving the device.

Install

Up and running in a minute

  1. Download the rln-lite build from the Releases page and unzip it.
  2. Double-click Rln-GUI.bat for the graphical app, or run rln-lite.exe for the command-line REPL.
  3. On first launch the file pickers open in the bundled examples folder — try use "demographics.csv" right away.

macOS and Linux builds run the same way. Prefer source? git clone, pip install -r requirements.txt, then python main.py --gui.

  1. Download rln-<version>-arm64-v8a-debug.apk from the Releases page.
  2. Copy it to your phone and open it; allow installs from unknown sources if prompted.
  3. On the first Open, grant All files access so Rln can read your data (Android 11+ requires this).

Quick tour

A 30-second taste

Type these in the Console — desktop REPL or the phone's Console tab. Bare file names resolve against the shipped examples folder.

use "demographics.csv", clear
summarize age income n_children
tabulate gender
histogram age

* larger-than-RAM mode — millions of rows, streamed
lrtm use "eu_firms_panel.parquet"
lrtm summarize

Watch

See it run

Data exploration and larger-than-RAM statistics — running natively on a phone.

New in 1.2.8

A color-coded data explorer, everywhere

The data browser now reads like a spreadsheet on every edition — numbers in blue, text in orange, missing values muted — using one shared scheme across the desktop GUI, the terminal browser, and the Android app.

Rln data browser — 420 obs × 4 vars
#    name        age   income
1    Chiara      48   152270
2    Eka         40   -12872
3    Satoshi     .    42747
4    Mwangi      35   98140

Preview Parquet before you load it

Run lrtm use "data.parquet" and the data browser instantly streams a live preview — no full load. Explore the whole dataset after lrtm collect.

Examples open by default

Both editions start in the bundled examples folder, so the sample data and scripts just open. Pin your own with set workdir "…" any time.

numbers negative strings missing

Capabilities

What works in each edition

A green tick means it works out of the box. The mobile estimation engine is a validated NumPy/SciPy backend — coefficients, standard errors, and common diagnostics match the desktop output.

CapabilityLite DesktopMobile (Android)
Read CSV, Stata .dta, Excel, JSON, dBase
Read & stream Parquet (LRTM)
Live Parquet preview before collect
Explore: describe, summarize, tabulate, correlate
Color-coded data explorer
Charts: histogram, scatter, line, twoway
Larger-than-RAM engine (polars)
Linear regression (regress)
Logit, probit, Poisson
Panel fixed effects (xtreg, fe)
Difference-in-differences (TWFE)
IV / 2SLS (ivregress)Desktop only
Advanced DiD (Callaway–Sant'Anna, …)TWFE fallback
Fuzzy merge (fuzzmerge)
Reproducible scripts + editor
NLP: summarize-text, translate, neuralOffline / Full tier

Highlights

Built for real research workflows

Read anything

CSV, Stata .dta, Excel, Parquet, JSON, dBase, R data — all editions.

Explore & describe

describe, summarize, tabulate, correlate, codebook.

Charts

histogram, scatter, line, twoway — rendered right in the app.

Larger-than-RAM

Stream millions of rows with polars — 2,000,000 summarized in ~0.04 s, even on a phone.

Econometrics

regress, logit, probit, poisson, panel FE, diff-in-diff, IV (desktop).

Fuzzy merge

Approximate-string joins that run on every edition via an automatic backend fallback.

Reproducible scripts

Multi-line .rln / .do scripts with syntax highlighting and an editor.

Examples folder

Sample data and scripts open by default; change the working folder any time.

In action

Screenshots

A full desktop session up top; the phone app shares the same design below — click any image to enlarge.

Rln desktop GUI streaming a 2-million-row Parquet with the script editor and console side by side
Desktop GUI — larger-than-RAM streaming of a 2,000,000-row Parquet, with the script editor and live console side by side.
Color-coded data browserData browser (color-coded)
Console explorationConsole exploration
Larger-than-RAM, 2 million rowsLRTM — 2M rows
Variables browserVariables browser
Plot histogramPlots
Script editorScript editor

Reference

Command quick reference

A taste of the language — over a hundred commands in total. Filter to find one.

use · importLoad a dataset (auto-detects format)
browseOpen the color-coded data explorer, or browse a file directly
describe · codebookInspect variables, types, and labels
summarize · tabulateSummary statistics and frequency tables
correlate · pwcorrCorrelation matrices
generate · replace · egenCreate and modify variables
recode · destring · encodeClean and convert variables
sort · gsort · collapseOrder and aggregate data
merge · append · fuzzmergeCombine datasets (exact or approximate)
histogram · scatter · twowayCharts rendered in the app
regress · ivregressLinear and instrumental-variables models
logit · probit · poissonGeneralized linear models
xtregPanel fixed / random effects
didregressDifference-in-differences estimators
predict · margins · testPostestimation
lrtm …Larger-than-RAM streaming with polars (preview before collect)
do · doeditRun and edit reproducible scripts
set workdirPin the default working folder
ssc installAdd optional backends (source installs)
helpFull command list and per-command help
No commands match that filter.