Lecture 3

Question 2

What are the three criteria for a tidy dataset?

Answer 2

1. Each variable has its own column. 2. Each observation has its own row. 3. Each value has its own cell.

Question 3

What are common signs of untidy data?

Answer 3

Column headers are values, not variable names.

Multiple variables stored in one column.

Variables stored across both rows and columns.

A single observational unit stored across multiple tables.

Question 4

What is the function used to convert wide data to long data?

Answer 4

melt()

Question 5

What is the function used to convert long data to wide data?

Answer 5

dcast()

Question 6

How do you split a column into multiple columns?

Answer 6

Use separate(), e.g., separate(data, col = “proportion”, into = c(“votes”, “total_votes”)).

Question 7

How do you combine multiple columns into one?

Answer 7

Use unite(), e.g., unite(data, col = “candidate”, name, surname, sep = “ “).

Question 8

What function is used to concatenate multiple tables?

Answer 8

rbindlist()

Question 9

What are the four types of merges in data.table?

Answer 9

Inner Merge: Only matching rows from both tables.

Outer Merge: All rows from both tables with NAs for missing values.

Left Merge: All rows from the first table, with NAs for non-matching rows in the second.

Right Merge: All rows from the second table, with NAs for non-matching rows in the first.

Question 10

How do you perform an inner merge in data.table?

Answer 10

merge(table1, table2, by = “column”, all = FALSE)

Question 11

How do you merge two tables by multiple columns?

Answer 11

merge(table1, table2, by = c(“col1”, “col2”))

Question 12

Why is there no single tidy representation of a dataset?

Answer 12

The tidy representation depends on the observation and the goal of the analysis.

Question 13

What is the difference between back-end and front-end data needs?

Answer 13

Back-end: Data is normalized to avoid redundancy.

Front-end: Data may be combined for easier analysis, even with some redundancy.