Andrew Worthy

Andrew Worthy

PhD Researcher & Data Scientist Python · ML · Signal Processing

MonStim Plotter — Project Write-up

Published:

MonStim Plotter is a desktop application for automated batch generation of publication-quality EMG trace figures. Researchers configure a styling template once and then point the tool at a directory of recordings to produce consistent, print-ready figures across an experiment.
Python PyQt6 Matplotlib Batch Processing YAML Configuration PDF / SVG Export

The Problem

Every paper our lab produces requires dozens of EMG trace figures: individual recording sweeps overlaid across stimulus intensities, annotated with peak markers, and styled to match journal requirements. Before this tool existed, each figure was produced by hand — researcher by researcher, with no shared standard.

The consequences were predictable:

  • Inconsistency: figures for the same experiment differed in font size, lineweight, axis limits, and color between researchers
  • Non-reproducibility: updating a figure for a revision (e.g., changing axis limits or color scheme) meant recreating it from scratch
  • Time cost: generating the figure set for a typical manuscript section took half a day

My Approach

The key design insight was to treat figure generation as a data pipeline with a configuration layer. The data (which sweeps to include, how to normalize) and the presentation (colors, font sizes, layout) should be fully separated — so changing the appearance never requires touching the data logic.

I built MonStim Plotter as a companion to MonStim Analyzer: it reads the same session files, applies a researcher-defined configuration, and outputs a complete figure set.

Technical Implementation

Modular configuration system

Researchers define a YAML configuration file specifying:

  • Which channels and sweep ranges to include
  • Figure dimensions and DPI for target publication/journal
  • Color palette (per-channel or condition-based)
  • Font sizes, lineweights, axis label formats
  • Annotation style for peak markers

The config file is version-controlled alongside the data, so the complete recipe for any figure in a paper is archived and reproducible.

Separation of data and rendering

The application has two fully independent module layers: a data layer (inherits MonStim Analyzer’s session object model) and a rendering layer (Matplotlib-based). Swapping a colormap or font size requires changing only the config — the data layer is untouched. This separation made iterating on figure styles fast and safe.

Batch processing

Given a session directory and a config file, MonStim Plotter processes every recording and exports figures in a single pass. Output formats — PNG (screen-optimized), SVG (vector, editable in Illustrator/Inkscape), and PDF (print-ready) — are generated simultaneously, so researchers have the right format for any downstream use.

Interactive preview

Before committing to a full batch run, researchers can preview any individual figure in the application window, adjust config parameters live, and confirm the output looks correct — preventing wasted batch jobs.

Results & Impact

  • Figure generation for a typical manuscript: reduced from approximately 2–4 hours of manual work to under 10 minutes of automated batch processing
  • Adopted as the lab standard for EMG trace figures, ensuring consistent styling and formatting
  • Reproducibility: the config file for each figure is archived with the data, enabling exact recreation
AEWorthy/MonStim-Plotter