Week 6 - Different Data formats

Question 1

What kind of data is .sav?

Answer 1

Binary format data for statistical packages such as SPSS.

Question 2

How should we handle missing values?

Answer 2

Either replace or remove.

Question 3

What is the difference between read_csv and read.csv

Answer 3

Read_csv is quicker and shows the progress. Automatically parses different types of information.

Question 4

What kind of format is climate data stored as?

Answer 4

netCDF (network common data form)

Question 5

What is JSON?

Answer 5

JSON essentially replaces XML. It is commonly used for web and is language independent.

Question 6

What should we do with missings for a plot?

Answer 6

Keep them in the plot, but make it obvious they are missing (i.e. in the border)

Question 7

How do we handle missings?

Answer 7

If there are only a few missings, drop them.
If many missings for a variable - drop it.
Few missings in many vars and cases - Impute values.

Question 8

What are the common ways to impute values?

Answer 8

Simple parametric: using mean or median
Simple non-parametric: find k nearest neighbours with complete values and average
Multiple imputation: using stat distribution e.g. normal model and simulate the missing.

Question 9

What is a contingency table?

Answer 9

Where you have two categorical variables and you count the occurrences.