M348

You will enjoy M348 if…

you love exploring interesting datasets
you want to solve practical problems by modelling
econometrics or data science sounds fun
you want amazing, free statistical software #rstats

Coursebooks

Three coursebooks (including both strands!) and a handbook (click to enlarge)

Computing Environment

R (free, open-source, first-class language 💚)
All software runs in browser - VLE with all packages preinstalled and ready to go.

Day-in-the-life of M348

load("mannaAsh.RData")
summary(mannaash)

Error in eval(expr, envir, enclos): object 'mannaash' not found

load("mannaAsh.RData")
summary(mannaAsh)

     treeID         diameter          height         side   
 Min.   :236.0   Min.   :0.1500   Min.   : 4.000   east:11  
 1st Qu.:247.2   1st Qu.:0.2000   1st Qu.: 7.000   west:31  
 Median :258.5   Median :0.2400   Median : 8.000            
 Mean   :257.8   Mean   :0.2424   Mean   : 8.024            
 3rd Qu.:268.8   3rd Qu.:0.2775   3rd Qu.: 9.000            
 Max.   :279.0   Max.   :0.3500   Max.   :10.000

Aha!

load("mannaAsh.RData")
side <- mannaAsh$side
plot(mannaAsh$diameter, mannaAsh$height, pch=as.numeric(side))
legend("bottomleft", title="side", legend=levels(side), col=5, 
pch = seq_along(levels(side)))

load("mannaAsh.RData")
heightSimple <- lm(height ~ diameter, data=mannaAsh)
heightParallel <- lm(height ~ side + diameter, data=mannaAsh)
anova(heightSimple, heightParallel)

Analysis of Variance Table

Model 1: height ~ diameter
Model 2: height ~ side + diameter
  Res.Df     RSS Df Sum of Sq    F    Pr(>F)    
1     40 105.055                                
2     39  55.293  1    49.763 35.1 6.594e-07 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

So the formal test result is significant - confirming our intuition. We should include “side” in the model, giving two parallel slopes.

Thanks for visiting!

The tutors and students of M348 are a wonderful mix of economists, computer scientists, mathematicians… You’ll love it!

via GIPHY