Draw me a picture
You will read a series of scenarios describing a dataset and communication intention.
For each scenario, imagine what a relevant data visualization could look like.
Explore as many ideas as possible for each scenario.
Tips:
- Identify the relevant variables and their relationship
- It may help to first imagine what the underlying data would look like
#1
A company claims their pill increases IQ.
- 20 participants are enrolled in your study. You measure their IQ before and after they take the pill.
- Make a plot showing the effectiveness of the pill.
#2
A company claims their pill increases IQ.
- 40 participants are enrolled in your study.
- You measure their IQ before and after they take the pill.
- 20 participants get the real pill
- 20 participants get a fake pill (placebo)
What type of claim would you like to make about this study? Make a plot showing that.
#3
A company claims their pill increases IQ.
1000 participants are enrolled in your study. You measure their IQ before and after they take the pill. The dosage of the ingredient in the pill is varied across participants.
What type of claim would you like to make about this study? Make a plot showing that.
#4
A company claims their pill increases IQ.
1000 participants are enrolled in your study. You measure their IQ before and after they take the pill. The company claims the pill is most effective for younger women.
Design a type of plot that the company might want to show.
#5
A company claims their pill increases IQ.
The company sold pills to 5 different universities. The number of students and their average IQ is different across each University. Some Universities enroll their students in multiple IQ increasing programs. From each student you know the age, gender, grades before the pill, grades after the pill, nationality, IQ before the pill, IQ after the pill and self-rating of well-being. Are the effects of the pill on IQ different across the 5 universities?
#6
Some customers complain that the pill has no effect on them. You wonder whether the pill is effective for some people but not others.
Using any data you collected until now, how would you test this hypothesis visually?
#7
The company wants to know the long-term effects of taking their IQ-increasing pill once.
20 participants are enrolled in your study. You measure their IQ before they take the pill. Then you repeat that IQ measure every day for the next week.
#8
A biochemist suggests that the effectiveness of the molecule depends on your hormonal state, which varies across the day. On the other hand, you believe that IQ measures themselves will vary depending on the time of day (sleepy people perform worse).
You have unlimited resources to test this hypothesis. But before you start, what would it look like if the pill is indeed more effective at certain times of the day?
#9
You are investigating a legal case against a company that claims their pill increases IQ. The data so far indicates that this is true for some people, but you suspect a methodological flaw: people were allowed to refuse the post-pill IQ test if they wanted to, and if so, they were excluded from the study (the “out” group).
Imagine different scenarios of what the data for the “out” group would look like if they had completed the post-pill IQ test.
Which of these scenarios did the company assume?
#10
You are the head of marketing for the company selling IQ-increasing pills. You want your clients to know the effect a given pill will have on them. The effects of the pills vary in intensity (how much IQ is increased), longevity (how the effect starts and wears off), and severity of side effects (headaches, jitters, etc.). The effects of each pill interact with features of the person such as their age, gender, and BMI category (small, medium, large).
What do you do?
#11
You measured 5 variables and want to know how they relate.
You measured 10 variables and want to know how they relate.
You measured 100 variables and want to know how they relate.
Share your drawings on Moodle.
