Load the Pokemon dataset

Pokemon dataset from https://www.openintro.org/stat/data/?data=pokemon.

Option 1, after downloading the csv file to the same directory as this script (PokemonGO.Rmd):

pokemon <- read.csv("pokemon.csv")

Option 2, downloading from the internet:

pokemon = read.csv("https://www.openintro.org/stat/data/pokemon.csv")

Linear models on Combat Power

pfit = lm(cp ~ species + hp + weight + height, data=pokemon)
pfit2 = lm(cp ~ (species + hp + weight + height)^2, data=pokemon) #all 2-way interactions
pfit3 = lm(cp ~ . - notes - name - attack_weak, data=pokemon)  # all available predictors except for notes, names, and attack_weak.

dplyr introduction

Help at https://www.rstudio.com/wp-content/uploads/2015/02/data-wrangling-cheatsheet.pdf

Create a dataset that excludes all columns whose name ends in “new” (notice the - before ends_with():

library(dplyr)
## 
## Attaching package: 'dplyr'
## The following objects are masked from 'package:stats':
## 
##     filter, lag
## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union
pokemon2 = select(pokemon, -ends_with("new"))

Exclude columns whose name ends in new, then keep only rows where the species is “Pidgey”:

pokemon2 = pokemon %>% 
  select(-ends_with("new")) %>%
  filter(species == "Pidgey")
summary(pokemon$species)
## Caterpie    Eevee   Pidgey   Weedle 
##       10        6       39       20
summary(pokemon2$species)
## Caterpie    Eevee   Pidgey   Weedle 
##        0        0       39        0

ggplot2 introduction

Help at https://www.rstudio.com/wp-content/uploads/2015/03/ggplot2-cheatsheet.pdf

A blank canvas

ggplot2 plots are built up piece by piece. The following creates a blank plot, and store all the data to make that plot and build on it in the myplot object:

library(ggplot2)
myplot = ggplot(aes(cp, cp_new), data=pokemon)

A scatterplot

To add a scatterplot with colors and a legend, just add a geom_point() call. We could save this to a new object, like myplot2, but in this example we won’t create any new object, just make a scatterplot:

myplot + geom_point(aes(color=species))

A scatterplot plus regression lines

Try doing the following one by one, just adding new things to the existing plot. These functions are all documented in the ggplot2 cheat sheet. You can also try skipping some of the lines. Note that nothing will happen until there is a line that doesn’t end in “+”

ggplot(aes(cp, cp_new), data=pokemon) +
  geom_point(aes(color=species)) +  #scatterplot
  geom_smooth(method="lm") +  #linear regression line and confidence bands
  theme_bw()  #get rid of the grey background

Boxplot

Now let’s make a boxplot:

ggplot(aes(species, cp), data=pokemon) +
  geom_boxplot(fill="grey") +
  ggtitle("Combat Power by Species") +
  xlab("Species") + 
  ylab("Combat Power") +
  theme_bw()

Try other kinds of plots, geom_violin and geom_dotplot…