Graphical Displays

Question 1

Explain bivariate exploration tools for numeric x numeric variables

Answer 1

Numeric x Numeric
* Descriptive statistics: Correlations
* Visual displays: Scatterplots
* Observations: Any noticable relationships? Ex. monotonic or non-linear?

Question 2

Explain bivariate exploration tools for numeric x categorical variables

Answer 2

Numeric x Categorical
* Descriptive statistics: Mean/median split by categorical variable
* Visual displays: Split boxplots, histograms (stacked or dodged)
* Observations: Any sizable differences in the means/medians among the factor levels?

Question 3

Explain bivariate exploration tools for categorical x categorical variables

Answer 3

Categorical x Categorical
* Descriptive statistics: 2-way frequency table
* Visual displays: Bar charts (stacked, dodged, or filled)
* Observations: Any sizable differences in the proportions among the factor levels?

Question 4

What are graphical displays to detect interactions between numeric x categorical variables?

Answer 4

Numeric x Categorical
* Numeric target: scatterplot colored by categorical predictor
* Categorical target: boxplot for numeric predictor split by target and faceted by categorical predictor

Question 5

What are graphical displays to detect interactions between categorical x categorical variables?

Answer 5

Categorical x Categorical
* Numeric target: boxplot for target split by one predictor and faceted by the other predictor
* Categorical target: bar chart for one predictor filled by target and faceted by the other predictor

Question 6

What are graphical displays to detect interactions between numeric x numeric variables?

Answer 6

Numeric x Numeric
* Bin one of the predictors (for example, cut it into several ranges), or try a decision tree

Question 7

Explain univariate exploration tools for numeric variables

Answer 7

Numeric Variables
* Descriptive statistics: Mean, median, variance, min., max.
* Visual displays: Histograms, boxplots
* Observations: Are there any (right) skews? Any outliers?

Question 8

Explain univariate exploration tools for categorical variables

Answer 8

Categorical Variables
* Descriptive statistics: Class frequencies
* Visual displays: Bar charts
* Observations: Which levels are most common? Any sparse levels? For binary targets, is there a presence of imbalance?