This exercise practices what you learned in the R crash course of Session 2: reading data, working with data frames, summary statistics, correlation (Session 1) and plots. Work in this file: write your code in the chunks marked # Your code here, answer the questions in the text below them, and knit the file to HTML when you are done.

The file BMI-steps.csv in this folder contains data on 1786 people: the average number of steps they walk per day, their body mass index (bmi), and their gender. A common hypothesis is that people who walk more have a lower BMI. Let’s see what these data say.

1. Load and inspect the data

Read the file BMI-steps.csv into a data frame called df. How many rows and columns does it have? Print the first rows and a summary of each column.

df <- read.csv("BMI-steps.csv")
# Your code here

2. Summary statistics

Calculate the mean, median and standard deviation of steps and of BMI. Then calculate the mean number of steps and the mean BMI separately for women and for men.

Hint: you can select the rows of one group like in the crash course, e.g. df[df$gender == "woman", ].

# Your code here

Plot a histogram of steps and a histogram of BMI. Do the distributions look like you expected?

# Your code here

3. Correlation

Calculate the Pearson correlation coefficient between steps and BMI with cor(). Is it positive or negative? Is it strong or weak?

# Your code here

Now calculate the correlation separately for women and for men. Is it different?

# Your code here

What would you conclude about the hypothesis from these numbers alone?

Your answer here.

4. Look at the data

Make a scatter plot with steps on the x-axis and BMI on the y-axis. Colour the points by gender, for example with col = ifelse(df$gender == "woman", "red", "blue") inside plot().

# Your code here

Do you notice anything else? What does this tell you about computing statistics before looking at the data? Think back to the Datasaurus dozen at the end of Session 1.

Your answer here.