Seminar 1

Question 1

Name three types of central tendency (averages)

Answer 1

Mean, Median, Mode

Question 2

Name three ways to measure the spread of a data set

Answer 2

Standard Deviation, Range, Variance

Question 3

Which type of central tendency can you use for categorical variables?

Answer 3

The mode (the most frequent category)

Question 4

You are comparing average customer satisfaction between two products. Can you rely on descriptive statistics to generalise your findings?

Answer 4

No. we can never generalise findings using descriptive statistics. This is particularly true when talking about continuous variables (i.e., real numbers).

Question 5

Name the three stages of data analytics

Answer 5

Descriptive analytics, Predictive analytics, Prescriptive analytics

Question 6

Describe Descriptive analytics.

Answer 6

Basically inferential statistics. Analysing historical data (events that have already taken place)

Question 7

Describe Predictive analytics

Answer 7

Building mathematical, computational, and statistical models to make predictions using existing data. In most cases we look at making numerical (regression) or categorical (classification) predictions.

Question 8

Describe Prescriptive analytics

Answer 8

Building data-driven solutions to control, or change the outcome of an event (e.g., helping customers build credit, or directing salespeople on best targets for advertising merch).

Question 9

A team of medical researchers is testing a new drug on hypertensive patients. They hypothesise that the drug will alter systolic blood pressure. Write out the null and alternative hypotheses.

Answer 9

Null hypothesis (there is no difference in the means on a population level, i.e., significant difference. Denoted H0).

Alternative hypothesis (there is a difference in the means. Denoted H1)

H0: There is no effect of drug on systolic BP

H1: There is an effect of drug on systolic BP

Question 10

Can you generalise means and use them to gain conclusions?

Answer 10

No. We cannot generalise using the means. If there is a large difference we need to un some formal tests first.

Question 11

What is random sampling?

Answer 11

Randomly sample from the population. Simple, yet risks overrepresentation of certain groups, unequal group sizes.

Question 12

What is stratified sampling?

Answer 12

Allows for control of group sizes by sampling based on said groups (e.g., sex, profession, etc.). Needs careful planning as frequencies may lead to biased representation of otherwise smaller groups by overrepresenting their importance.

Question 13

What is clustered sampling?

Answer 13

Usually based on geography and proximity. E.g., sampling from a local hospital, rather than all hospitals.

Question 14

What is systematic sampling?

Answer 14

Taking every kth member. Especially useful in industrial domains where order may matter.

Question 15

What is the difference between a parameter and a statistic?

Answer 15

Parameters are summaries of population data. Statistics are summaries of sample data. In many places there are used interchangeably (e.g., in data science).

Question 16

Name two types of qualitative data.

Answer 16

Categorical , Ordinal

Question 17

Describe Categorical data

Answer 17

No particular order to them. E.g., countries, eye colours, sex, etc.

Question 18

Describe Ordinal data

Answer 18

An order has been applied. E.g., Likert-scaled questions from surveys, countries by population size, products by amount sold, etc.

Question 19

Name two types of quantitative data

Answer 19

Discrete, Continuous

Question 20

Describe Discrete data

Answer 20

Whole numbers (integers). E.g., number of emails sent.

Question 21

Describe Continuous data

Answer 21

Real numbers, such as temperature, weight, height, etc. There are an infinite number of real numbers even between two numbers such as 0 and 1.