Plots types are as difficult to classify as the data they represent. There are cases in which the selection of a visualization is clear cut, but more often than not, the selection of the appropriate way to express our data depends on the context we want it to appear in. This can be as wide as the media we are gonna use it for (live demo, print, etcetera), or the field we work in (network theorists prefer graph visualizations, whilst genetic analysts prefer heatmaps). For these reasons, it is difficult to classify the uses for which some plots are better than others, but one "general" way to do so is by classifying them by the function they perform.
In the following module, we will show some of the most common ways to display information according to the underlying data they hold, and the use we want to give to it.
- Data type: data counts varying over time.
- Some options: joined scatter plots, stream chart (stacked area), dygraph
- Data type: frequencies or ratios
- Some options: histogram, smooth histogram, box-whisker chart, violin plot
- Data type: {x,y} or {x,y,z} coordinates
- Some options: scatter with histogram, bubble chart
- Data type: matrices that represent transitions between states.
- Some options: networks, chord diagrams, heatmaps
- Data type: two or more independent variables, and two or more dependent variables
- Some options: contour plot, density plot
- Data type: data with high dimensionality (for classification, from factorial experiments).
- Some options: 4d contour plots, parallel line plots
- Data type: geodata
- Some options: map, globe
- Data type: hierarchical
- Some options: treemap, networks, dendrogram, scatter plots
