# ferdo R Lab | Lesson 5 v1.0 | Student independent tasks
# All records are invented. Keep data/class_synthetic.csv untouched.
stopifnot(file.exists("data/class_synthetic.csv"))
d <- read.csv("data/class_synthetic.csv",
              na.strings=c("","NA"),stringsAsFactors=FALSE)
ok <- complete.cases(d[,c("study_hours","quiz_score")])
paired <- d[ok, ,drop=FALSE]

# TODO 1: Count complete pairs and identify excluded fictional IDs.

# TODO 2: Create a scatterplot of paired study hours and quiz score.

# TODO 3: Use cor() and explain direction/strength without causal claims.

# TODO 4: Fit a simple linear model with lm(); inspect coef().
# What are the units of the slope?

# TODO 5: Plot the fitted line on the scatterplot with abline().

# TODO 6: Obtain model R-squared and explain what it DOES NOT prove.

# TODO 7: Predict quiz score for 4 study hours.
# Then predict for 10 hours: why is extrapolation dangerous?

# TODO 8: Calculate residuals and identify which residual is largest
# in absolute value; report the corresponding fictional student ID.

# TODO 9: In a COPY ONLY, replace S08's quiz score with 50.
# Recompute cor() and discuss sensitivity. Do not edit the original CSV.

# TODO 10: Write a short English summary: n, correlation direction,
# slope in points/hour, and two substantive limitations.