Data Visualisation And Publishing
April 23, 2026 · View on GitHub
Create static and interactive data visualisations for reports, repos, and data storytelling — prioritising a curated inventory of open-source tools.
This plugin opinionates the tool choice. Instead of reaching for whatever library the model happens to recall, it selects from a maintained inventory (Matplotlib, Bokeh, Chart.js, ECharts, D3, visx, Vizzu, VChart, Plotly Dash, Lightweight Charts, fl_chart, Constellation, DataWarrior, Iris, QuantInvestStrats, react-globe.gl, and others) and matches the tool to presentation purpose, data shape, runtime, and audience.
Skills
- tool-inventory — canonical catalogue of preferred libraries grouped by use case (scientific, web, app, animated, Python, geospatial).
- choose-tool — decision protocol that maps presentation purpose + data shape + runtime to a tool in the inventory.
- setup-environment — scaffolds a reproducible dataviz project (Python / web / Flutter / desktop) with a working "hello chart" example.
- data-storytelling — narrative-driven visualisation design: headline-first charts, annotation, animated transitions, interactive elements that earn their place.
Commands
/dataviz-pick— pick the right tool from the inventory for a given task./dataviz-setup— scaffold a dataviz environment using the chosen tool./dataviz-story— design and build a data-storytelling piece.
Philosophy
- Pick from the inventory first. Only reach outside if nothing fits, and justify it.
- Match the medium. Static PDF figure, web embed, dashboard, animated story, and geospatial all want different tools.
- Storytelling ≠ dashboards. If the goal is persuasion, build one chart one claim; if the goal is exploration, build a dashboard.
- Reproducible by default. Pin deps, commit lockfiles, keep data/code/output separate.
Installation
claude plugins install data-visualisation-and-publishing@danielrosehill
License
MIT