lecture three: intro to scatterplots and correlation

benefits of scatterplot?

• doubles info available from heat map

• improves ability to check for possible relationships

what is correlation?

• measure of empirical relationship

• analyses produce r value (THRESHOLD OF 0.3 OR HIGHER)

• correlation DOES NOT mean causation

what is the importance of scales?

• smts use non-zero starting points BUT can make small differences seem large

• “breaks” help outliers appear to be less extreme

• need to have correct increments

• CHECK AXES FOR BIAS

dangers of averages/rates?

leave out details of various data points that caused that result or time spans

average caution:

KNOW DIFFERENCE BTWN “TYPICAL” AND “MEAN”

what is current year value?

hide inflation effects (face value)

what is the fixed year value?

includes inflation effects (shows real buying power)

wealth and education in data maps

• states have wealthier and poorer residents

• taxes provide the largest share of gov revenue

• a state with poorer residents can’t provide same services as state with wealthier residents

• the lack of an educated public may keep a state poor as better paying jobs go elsewhere

• increases in education levels

• state per capita income in comparison to education suggests a relationship

scatterplots and correlations

• show positive, strong correlation between per capita income and percent high school completion

• relationship has changed over time

spurious relationships

• correlation DOES NOT mean causation

• often from misleadings from population - never compare the states and ignore population