From Data to Viz
Yan Holtz and Conor Healy
A decision tree for choosing a chart: start from the format of your data and follow it to the chart types that fit, with their common pitfalls.
What you will find
- A decision tree by data format: numeric, categorical, numeric and categorical, maps, networks and time series
- A caveats section on common mistakes, such as too many slices in a pie or dual axes
- Links to R and Python code for each chart
What we can learn from it
- 01Start from the shape of the data, then choose among the charts that fit it
- 02Every chart comes with its pitfalls; know them before you publish
Chart types you will meet there
Violin plot
Mirrors a smoothed distribution around a centre line, showing the full shape of each group.
Density plot
A smoothed histogram: a continuous curve showing where values concentrate.
Dendrogram
A tree with all leaves aligned, typically showing how items merge into clusters.
Correlation matrix
Shows the correlation between every pair of variables, coloured from strongly negative to strongly positive.
Similar sites
The Data Visualisation Catalogue
A reference library of chart types: one page per chart explaining what it shows, how to read it and which tools can make it.
Data Viz Project
A Copenhagen agency’s catalogue of chart types, browsable four ways: by family, by the data you have, by what you want it to do, and by the shape it makes.
FT Visual Vocabulary
The Financial Times’ one-page guide to choosing a chart, grouped by the relationship in the data you want to show.