// the find
holtzy/data_to_viz
Leading to the dataviz you need
data-to-viz is a static reference site built with R Markdown/knitr that walks you from the shape of your data (one number, two categories, a network, etc.) through a decision tree to appropriate chart types, then links out to worked examples and common pitfalls. It's a companion to the R Graph Gallery, aimed at anyone picking a chart type rather than developers looking for a library to install.
The decision tree is organized by input data shape rather than by chart name, which is the right way to frame the problem — you start from what you have, not from a list of 40 chart types you have to already know to search for. The caveats section (cut y-axis, dual axis, 3D pie, mishandled error bars) gives concrete failure modes with explanations, which makes it a useful thing to link a colleague to instead of re-explaining why their chart is misleading. Each recommended chart type ties back to actual R code in the sister Graph Gallery repo, so the advice isn't just theoretical.
This isn't a library or tool — there's nothing to install or import, it's only useful by visiting the hosted site, so as a GitHub repo it's mostly of interest if you want to fork the content or reuse the decision-tree structure. The repo has several full copies of d3 (v3, v4), bootstrap, and plotly vendored under caveat/libs from knitr's self-contained HTML output, which bloats the checkout and makes the source hard to actually read or extend. The README still describes the project as 'beta' and points feedback to a Twitter handle, and there's no sign of active content updates — the value here is whatever content already exists, not an evolving resource. Everything is R-only, so if your stack is Python or JS the code examples don't transfer, only the chart-choice logic does.