Week in Review: Pie Charts and Maps, Reproducibility, and Social versus Physical Science
Graphs from the week
- Interesting graphs of the cost of the war on terror.
- Jon Peltier looks at pie charts showing who is blamed for the mess in Washington. The Monkey Cage presents a pie chart on grading schools. Is a pie chart ever useful? Never!
- Maps with geographic data make amazing graphs. One map this week looks at the distance to the nearest McDonald's in the US and another replicates the figure for the UK. Another map graphs world population densities by income.
- Earlier this week, I added a post about colors in graphs and had a discussion with Jon Peltier in the comments section. Georgette Asherman offers some thoughts on how use colors effectively. Some R advice on converting values to colors.
- Changes in militancy of religions across time. Worst graph of the year?
Reproducibility of research came up several times in several forms.
- An article in Nature suggests the 50% of published studies might be wrong.
- Andrew Gelman kicked things off with a post about the statistical significance filter, and following up by commenting on type M errors in the lab and featuring a post about data management errors.
- Tyler Cowen and Bruce Booth offer their takes. Observations Epidemiology presents some thoughts as well.
- I offer some thoughts about how to encourage reproducible findings.
- An article in Significance makes some suggestions on encouraging reproducibility, and I offer comments.
- The Economist brought up reproducibility as well and Revolutions gave their take on it.
An interesting discussion also emerged on the differences between the hard and social sciences, especially their public perceptions.