7.1 Visualizing a single distribution

• Histograms are useful in visualizing distributions

• Basically turning a quantitative variable into a qualitative variable via binning

• Choice of how large each bin should be affects interpretation

• Density plots are similar to histograms but have important caveats:

  • The kernel choice (e.g. Gaussian, rectangular) should make sense given your knowledge of the distribution

  • May be misleading for discrete quantitative variables (e.g. counts)

  • Suggests data exists where none exists