ANIMATED VERSION NOTES
Requirements:
  Creates illusion of fluid motion
  Animation is meaningful (not frivolous)
  Animation parameters/variables are meaningfully named
  Control structures are used to enable animation
  Sketch uses linear transf. fxns to create interaction/animation
  Identify key-frames and use tweening to connect them.
  Use a form of randomness to make the output more interesting
    without losing meaning.
  Style the output according to a set of established aesthetic practices.
  The output and the code have clear meaning to a first-time user.
  The output and the code inspire awe, wonder, and/or action.

-------------------------------------------------------------------------------------------

INTERACTIVE VERSION NOTES
(scroll down for the animated version notes)

Create a scene controller variable to track which scene we're drawing

Create an event listener for:
  LEFT CLICK
  RIGHT ARROW KEY INITIAL PRESS
To increment the scene controller, modulo the total number of scenes

Create an event listener for:
  LEFT ARROW KEY INITIAL PRESS
To decrement the scene controller, modulo the total number of scenes

--Coding problem with tracking initial presses: You will need to implement a monostable circuit (cf. the madlib ex.).

Put a bunch of the repeated stuff into functions so it's easier to keep a hold of:
  DRAW AXES
  DRAW DATASET 1, DATASET 2
  DRAW REGRESSION 1, REGRESSION 2, COMBINED REGRESSION
  WRITE TEXT INTO BOTTOM OF CANVAS

Plop each scene into a function that dictates what to draw

SCENE #0:
  Blank plot
  Text that gives name of phenomenon, explains user controls (click, arrow keys)

SCENE #1:
  Set-up text
  You are tracking two variables: some independent variable x that influences the dependent variable y
    Say, as an example, we're looking at basketball players around the same age
      We let x = inches above 5 ft. and y = no. of free throws made out of 10
    
SCENE #2:
  You have two datasets: circles and squares
    These form two parts of the same population
  Draw circles and squares on the plot, in the centre
  
SCENE #3:
  Suppose the circles make a positive correlation
  SUPPRESS AXES
    The more x increases, the same thing follows for y
  Draw circles in a line with a regression

SCENE #4:
  Also suppose the same is true for the squares
  SUPPRESS AXES
  Draw squares in a line with a regression
  
SCENE #5:
  What do you think will happen if we plot both datasets at once?
  DRAW AXES
  Draw question marks on plot

SCENE #6:
  Did you think there would still be a positive correlation?
  In this case, there isn't! The combined datasets have a negative correlation.
  Draw both datasets with the combined regression

SCENE #7:
  Of course, if the two datasets happened to be positioned differently, there could still be a positive correlation.
  Draw moved datasets with positive combined regression

SCENE #8:
  But you can't tell which way the combined trend will go just by looking at the datasets invidually.
  Bring back the drawing from scene no. 6

SCENE #9:
  Sometimes when multiple groups of data have the same trend, combining the data causes the trend to reverse.
  This phenomenon is called {Simpson's paradox}{bold}.
  Draw scene no. 6 plus individual regressions

SCENE #10:
  When this occurs, there isn't a simple, direct correlation between your independent and dependent variables.
    Confounding variables may be at play.
      So in our example, maybe the basketball players had different coaching quality.
  Click the diagram or press the right arrow key to restart.
  Keep same drawing from scene no. 10