These are the studies and readings that we discussed in class regarding visual encoding of your data:
- Stephen Few on All Things Circular (the lesson: beware of circles!)
- Cleveland and McGill: The first to do an empirical study of graphical perceptions and a short post about the findings, from Nathan Yau on FlowingData
- An interesting dataset called Anscombe’s Quartet that is often cited in data viz talks–it has nearly identical statistical properties, but look entirely different when graphed.
Summary findings of encodings, from most accurate to least accurate:
- Position
- Length
- Angle
- Area
- Density and color saturation
- Color hue
Know some of the common chart types:
- Bar charts: Trends for categories
- Line charts: Trends for continuous series/continuous changes between x-axis (time series)
- Scatter plot: correlation
- Bubble plot: scatter plot + additional variable
- Pie chart: show proportions
- Area charts/stacked graphs: proportions