Statistics 101 lecture outline: Descriptive Statistics 1

Statistics 101: Descriptive statistics 1

Administrivia:
Frequency: histograms, Pareto
- relative frequency = Area = base * height
- Too narrow bins look noisy
- Too wide bins lose shape
- Impossible to "correctly" choose bin widths
quick and dirty histogram = stem and leaf
Histogram properties
- center / width
- mode = maximum point of histogram
- bimodal = oops--more than one (local) maximum
- skewedness = asymetric
Warping histograms
- Why might you want to have uneven bin sizes?
- Make a histogram of the following data:
  10 20 20 30 30 30 40 40 50
- Now via changing bin sizes this data can be made to have a "U" shape!
- A look at some data
  - payments made to CEO's of top companies
  - What is the center / width?
  - Thought question: Do we really want one person's pay to determine spread?
  - Play with the "hand" to change the bin widths
  - Try logging the data
last updated: $Date: 2006-04-05 17:38:25 -0400 (Wed, 05 Apr 2006) $