All

Question 1

What is Exploratory Data Analysis?

Answer 1

Techniques for summarising, visualising and reviewing data; the first step to analysing data

Question 2

What is tabular data?

Answer 2

Each record or observation represents a set of measurements of a single object or event

Question 3

What are the four things you want visualisation to show?

Answer 3

Distribution, Relationship, Composition, Comparison

Question 4

What is Distribution?

Answer 4

How a variable or variables in the dataset distribute over a range of possible values

Question 5

What is Relationship?

Answer 5

How the values of multiple variables in the dataset relate

Question 6

What is Composition?

Answer 6

How the dataset breaks down into subgroups

Question 7

What is Comparison?

Answer 7

How trends in multiple variables or datasets compare

Question 8

Why rescale a graph?

Answer 8

To increase visibility, and to find a ‘law’ (find a straight line)

Question 9

What is Data Wrangling?

Answer 9

Exploring and transforming data to make valuable insights

Question 10

What are the steps of Data Wrangling?

Answer 10

Obtain, Understand, Explore, Transform, Augment, Visualise

Question 11

What are the types of missingness?

Answer 11

Missing Completely at Random, Missing at Random, Not Missing at Random

Question 12

What is Missing Completely at Random?

Answer 12

The probability that the feature is missing is independent of the value of any other features

Question 13

What is Missing at Random?

Answer 13

The probability that the feature is missing is independent of the feature but can be affected by the values of other features

Question 14

What is Not Missing at Random?

Answer 14

The probability that the feature is missing can be dependent on the value of the feature

Question 15

How can you deal with Missing Completely at Random?

Answer 15

Only use complete data

Question 16

How can you deal with Missing at Random?

Answer 16

You can try to predict the values of the missing values. Deleting these values would be biased

Question 17

How can you deal with Not Missing at Random?

Answer 17

You cannot do much against this

Question 18

What is faceting?

Answer 18

Apply the same analysis to many comparable subsets of data, then put them side by side

Question 19

What are the two kinds of variables and how are they classified?

Answer 19

Identifier variables are the variables that we set up. Measurement variables are the variables we measure

Question 20

What is a general strategy for working with larger data sets?

Answer 20

Split the problem into smaller pieces, work on each piece individually, recombine the pieces. This is known as split-apply-combine

Question 21

What is Visual Encoding?

Answer 21

The way in which data is mapped into visual structures

Question 22

What is Visual Perception?

Answer 22

The ability to interpret the surrounding environment by processing information

Question 23

What three things should you consider when encoding?

Answer 23

Importance, Expressiveness and Consistency

Question 24

What is Pre-Attentive Processing?

Answer 24

The subconscious accumulation of information from the environment

Question 25

What are the four types of data?

Answer 25

Nominal, Ordinal, Discrete, Continuous

Question 26

What is Nominal data?

Answer 26

Named Categories

Question 27

What is Ordinal data?

Answer 27

Categories with an implied order

Question 28

What is Discrete data?

Answer 28

Only particular numbers

Question 29

What is Continuous data?

Answer 29

Any numerical value

Question 30

What is a Contingency Table?

Answer 30

A table of counts for cases, rows and columns are labelled with categorical variables, while the cell values are counts

Question 31

What are the two types of study?

Answer 31

Observational study, using existing observations and plentiful data, versus experimental study, where you create a specific experiment to gather the data for the study

Question 32

What are Confounders?

Answer 32

A confounder is a variable that causes changes in both the identifier and measurement variables

Question 33

Give the steps for hypothesis testing

Answer 33

Specify the Null (H0) and Alternate (H1) hypothesis, assume the null hypothesis is true and use the data, calculate the p-value, if it is high assume null hypothesis, else assume alternate hypothesis

Question 34

What are the two types of error?

Answer 34

Type 1 error, H0 is rejected when in reality it is true, or type 2 error, H0 is not rejected when in reality it is false

Question 35

What is the p value?

Answer 35

P-value is the lowest value at which the null hypothesis is rejected

Question 36

What is A/B Testing?

Answer 36

Randomized controlled trials used by companies

Question 37

What is Dimensionality?

Answer 37

The number of measurements available for each example in a dataset

Question 38

What is a Multivariate visualisation?

Answer 38

Visualisation of datasets that have more than three variables

Question 39

Why reduce dimensionality?

Answer 39

To reduce strain on computers and allow use by humans

Question 40

What are the two types of reducing dimensionality in a non linear way?

Answer 40

Global method assumes that all pairwise distances are of equal importance, while the local method assumes that only the local distances are reliable

Question 41

What are the two types of reducing dimensionality in a linear way?

Answer 41

PCA (Principal Components Analysis) finds the directions that have the most variance, while MDS (Multi-Dimensional Scaling) arranges the points to minimise discrepancy