• Some of the lecture slides don't load on a web browser
• Current alternative for viewing: Download the HTML or the RMD for the moment and view locally.
• I'm in the process of sorting it out

• Time: 1 Hour
• Task: Complete a data analysis and answer questions (like the assignments, but short - similar to the in class exercises)
• Conditions:
  • Open book
  • Internet access
  • No contacting / messaging / communicating friends during exam (penalty will be the same as cheating in exam)

• An Rmd file +
• The rendered HTML file
• The data
• Zipped / compressed together
• In the document:
  • Write about the data - what is it, what are the variables, what sort of questions are you asking, what did you need to do clean the data, what cleaning is remaining?
  • (About 1-2 paragraphs of text.)

• "How many questions should we ask?"
  • About 3 - roughly one per person in your group
• "How will the project be marked?"
  • You will all mark the projects as you watch them
• You will need to present for 5 minutes max - use this to help frame how long your analysis should be.
• 5 minutes is short - but it is actually a pretty serious challenge!
• But also consider that the project will be marked by me as well - you won't need to cover everything

Many options!

• You can submit an Rmarkdown document + HTML file
• You present with a flexdashboard (demo)
• You could even present as a slide show (like the lecture slides)
• How to share - Ideally you can use netlify drag and drop (demo)

• Worth 50% of your final grade
• NOT Open Book
• Calculator: Yes - Allowed to take in "Only HP 10bII+ or Casio FX82 (any suffix) calculator permitted"
• One A4 page of notes, printed on both sides. This will be collected at the end of the examination.
• Hurdle requirement: You must get 40% on the exam to pass the course (also true of the final project).

• I'll now talk about the questions in the exam, and some of the concepts you need to be familiar with

• These concepts will help guide what you focus on in the lectures

• Disclaimer: This list is not exhaustive - these are to help give you a sense of what I'm thinking about for each of the questions in the exam.

• Variables
• Observations
• Tidy data

• Converting "messy" data to tidy data
• Code / key functions to use to convert data into "tidy" data
  • e.g, gather, spread, separate, etc.

• Left Join
• Why do joins of data
• Predict output of a join
• Sketching out code to summarise data from a join

• How the grammar of graphics produces a plot - identify plots produced by which code
• Understand the focus of a given graphic on the data - questions like:
  • "What and how does this graphic make us focus on what feature of the data?"
  • "What do you learn from a graphic?"

• Principles of tidy missing data
• Interpreting graphics of missing data
• Predict output of function on data with missing values
• Imputation - what is it, which methods are good / bad / better

• Write down an equation of a model from code output

• The formula $y=3x+5$ is a function with input $x$, and output $y$, when x is ___ , the output is ___

$$\widehat{heigh{t}_{in}}\sim 3.62+0.78Widt{h}_{in}$$

• slope: For each additional inch the painting is wider, the height is expected to be higher, on average, by 0.78 inches.

• Intercept: Paintings that are 0 inches wide are expected to be 3.62 inches high, on average.

• If a paining is 5 Inches wide, what is it's estimated height?

• How to make predictions from a fitted model
• Understand what makes predictions good and bad
• Measurements of model fit:
  • R2 - what it is - what values mean good/bad fit?
  • A good idea to bring in an equation of R2
• Residuals - What do we expect to see
• Centering variables
• Think about how you can improve fit of models to your data

• Why write functions
• How to write a function in R
• How to take existing code and turn it into a function
• Identify potential mistakes in provided code
• Understand what map does

• From an association / correlation matrix, which are most or least related?
• Understand how you can convert a numeric matrix into a binary association matrix
• Understand how a correlation matrix (or other association) of data relates to a provided network (let's do an example)

I've had a great time this semester

I've been seriously impressed with you all - you are all so bright!

Thanks for being engaged and curious, I'm really excited to see your projects!

Also a special thanks to everyone for dealing with the 8am lecture!