Evaluating Data Visualization

Overview

Questions

• TODO

Objectives

• interpret data visualizations to generate a narrative about unfamiliar data
  
• Compare and contrast interpretations of different visualizations from the same dataset
  
• Propose questions about choices made during the data visualization process to explore potential hidden features of the data

Develop Qs Evaluating a Whisker Plot

Come up with a set of questions to help you evaluate the clarity, usefulness, completeness, and objectivity of this whisker plot about Flu shots.

** EXAMPLE **

Some things to consider are: - n of participants

• visual display

• the statistical choices of the authors

• any comparisons or relationships among the plotted elements

Some Hints

sample responses: - are the data in each of the groups mutually exclusive?

• are there outliers excluded? how?

• how are the data combined to form each point & whisker plot?

• how are the CIs (confidence intervals) computed? [in this case, because the quantitative axis is % increase, CI is a little funky here as an estimate of error around the value, and not an estimate of variability within a group]

• are the number of subjects in each group equal (or approximately equal)?

• is there a grouping of interventions that suggest generalizations? (e.g. 1 text vs 2 texts, content of reminder, format of reminder)

• is there an effect of time of day that a reminder is sent/received (vs. relative to an appointment time)

• how robust are the estimates to the selection of the regression model?

• soooo, what’s the takeaway from this figure? 😄