A statistical signal-flow platform

Empiria

A suite of twenty-eight open-source modules across five plugins that turns the modular synthesizer's patch cables into a live, manipulable surface for statistics, data science, and computational social science. Every parameter is a knob; every intermediate estimator is a control-voltage signal; every module ships with a real-time visualization matched to the statistic it computes.

Download v2.0.0 Read the manual (PDF) Read the paper

What it is

Statistics education has been urged for two decades to center on simulation-based reasoning rather than closed-form normal-theory derivation, but the dominant computational vehicles — R scripts, Python notebooks, single-purpose web applets — still hide the data-generating process and the inferential apparatus behind opaque function calls.

Empiria takes the opposite approach. It exposes the entire empirical workflow as patch cables on a rack: parametric sampling → sample-window estimation → online OLS regression → exact t-tests → BCa bootstrap → autocorrelation diagnostics → deterministic record-and-replay → CSV export.

The patch is the curriculum. The wires drawn between modules externalize the inferential logic of a lesson in a way no static diagram can — and because every random draw is Mersenne-Twister-seeded and every closed-form quantity is computed by the standard numerical recipe (Lentz CF for the incomplete beta, BCa for the bootstrap, Sanger's rule for online PCA), the same patch is also a reproducible analog computer.

The five plugins

Each plugin is its own family of modules sharing a colored family-stripe under the header strip.

METHODS · 15 modules

Statistical workflow

Sample · Frame · Regress · Test · Boot · Lag · Code · Tab · Strata · Cohort · Factor · Seed · Tape · Gauge · Quantity.

POLIS · 6 modules

Agent-based social models

Cascade · Discourse · Pareto · Dilemma · Diffusion · Network. Granovetter cascades, opinion dynamics, wealth condensation, iterated dilemmas, small-world / scale-free networks.

EPI · 1 module

Network epidemiology

Outbreak — SIR-style epidemic spread on Watts-Strogatz, Erdős-Rényi, and Barabási-Albert graphs. "Flatten the curve" is one patch cable away.

SPACE · 3 modules

Spatial dynamics

Life · Schelling · Turing. Conway's Game of Life with rule presets, Schelling segregation on a 24 × 24 grid, Gray-Scott reaction-diffusion traversing spots, stripes, labyrinths.

DECISIONS · 3 modules

Behavioral economics & cognition

Prospect · Bandit · DDM. Kahneman-Tversky prospect theory, K-armed bandits with ε-greedy / UCB1 / Thompson sampling, Ratcliff drift-diffusion.

Deterministic seeding

Every random module wraps std::mt19937 with an explicit, on-panel integer seed. Same seed produces byte-identical results on macOS, Windows, and Linux.

Exact numerics

Two-tailed t uses Lentz continued fractions for the incomplete beta. χ² uses regularized incomplete gamma. Bootstrap CIs are bias-corrected and accelerated (Efron 1987) with on-panel z₀ and â diagnostics.

Portable artefacts

.vcv patches are plain JSON. Tape's right-click Export buffer to CSV bridges Empiria into R / Python / Julia. The patch is a citeable scientific artifact.

Install

Step 1

Get VCV Rack 2

Free download for macOS, Windows, Linux.

vcvrack.com/Rack →
Step 2

Download Empiria

Grab the latest .vcvplugin archives from GitHub releases.

github.com/kevisc/empiria/releases →
Step 3

Drop into Rack

Move the archive into Rack's plugin folder; restart Rack. Modules appear under the SHLabs brand.

Full manual (PDF) →

Cite Empiria

If you use Empiria in published research or coursework, please cite the companion paper.

@article{empiria2026,
  author  = {Schoenholzer, Kevin},
  title   = {Empiria: A Modular-Synthesizer Platform for
             Simulation-Based Statistics Education},
  journal = {Journal of Statistics and Data Science Education},
  year    = {2026},
  note    = {under review}
}

Free forever. Built by one person.

Empiria has no paid tier and no commercial sponsor. If it has saved you time, made your teaching land, or just made statistics legible to a student who needed it, consider supporting future development.

Support the work →