RickAndMortyAPI_EDA.Rmd

---
title: "Rick and Morty EDA"
author: "Julia Layne"
date: "4/19/2020"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

## Builing the Initial API Call


```{r}
#install.packages("httr")
#install.packages("jsonlite")
library(httr)
library(jsonlite)
library(plotly)

base <- "https://rickandmortyapi.com/api/"
endpoint <- "character/"
page <- '?page='
#25 pages, can I loop?
call1 <- paste(base,endpoint,page,"1", sep="")
#https://rickandmortyapi.com/api/character/?page=1
get_characters1 <- GET(call1)
get_characters1
```
# Pull out Text of Json response

```{r}
get_characters_text1 <- content(get_characters1, "text")
get_characters_text1
```

#Transform text block into data from using jsonlite
jsonlite recognises the text block as json and transforms it into a useable dataframe
```{r}
get_characters_json1 <- jsonlite::fromJSON(get_characters_text1, flatten = TRUE)
get_characters_json1
```

Specifically, we'll use the results to build our character dataset
```{r}
total <- get_characters_json1$results
total

```

We need to know how many pages we should iterate through
```{r}
pages <- get_characters_json1$info$pages
pages
```

## Looping calls

We'll loop through for every page and rbind the resulting response to our character dataset
```{r}
#[2,3,4,5.......30]
for (val in seq.int(2, pages, 1))
{
  call <- paste(base,endpoint,page,val, sep="")
  get_characters <- GET(call)
  get_characters_text <- content(get_characters, "text")
  get_characters_page <- jsonlite::fromJSON(get_characters_text, flatten = TRUE)
  
  
  total <-rbind(total, get_characters_page$results)
}
#summary(total)
```


# Defining 'Rick' and 'Morty'-ness
Using a simple string find through grepl to see if name contains one of these character names. 

```{r}
names <- c('test','Rick Sanchez','Morty Smith','Summer Smith')

rickOrMorty_name <- function(name) {
  rickOrMorty <- 'Neither'
  if(grepl('morty', name,ignore.case = TRUE)){rickOrMorty <-'Morty'}
  if(grepl('rick', name,ignore.case = TRUE)){rickOrMorty <-'Rick'}
  rickOrMorty 
}
#testing top dataset to confirm it is working
sapply(names, rickOrMorty_name,simplify = TRUE)
```

# Apply function over entire dataset to categorize all characters
```{r}
all_characters <- total %>% mutate(rickOrMorty = sapply(name, rickOrMorty_name,simplify = TRUE))
all_characters
```

# Subsetting data to only fields we care about
```{r}
#unnest(all_characters, .drop = NA, .id = episode, .sep = NULL,.preserve = NULL)

myvars <- c("name", "rickOrMorty", "status","species")
newdata <- all_characters[myvars]
newdata

```


## Graphs
```{r}
library(GGally)
newdata %>% select(rickOrMorty,status, species) %>% ggpairs(mapping = aes(color=species)) + ggtitle("Character by 'Rick or Morty', Health Status, and Species")
```

# Looking into Health Status
```{r}

p= newdata %>% 
  ggplot(aes(x = species, fill=status)) + 
  geom_histogram(stat="count", mapping = aes(color=status)) + ggtitle("Health Status by Species") + scale_fill_hue(l=80, c=12) + scale_color_hue(l=40, c=35)

ggplotly(p)

noHumans= filter(newdata, species != 'Human') %>% 
  ggplot(aes(x = species, fill=status)) + 
  geom_histogram(stat="count", mapping = aes(color=status)) + ggtitle("Health Status by Species (No Humans)") + scale_fill_hue(l=80, c=12) + scale_color_hue(l=40, c=35)

ggplotly(noHumans)

noHumanDf <-filter(newdata, species != 'Human')
noAliensNoHumansDf <- filter(noHumanDf, species != 'Alien')
noAliensNoHumansDf

filter(noHumanDf, species != 'Alien' & species != 'Human')

noHumansOrAliens= noAliensNoHumansDf %>% 
  ggplot(aes(x = species, fill=status)) + 
  geom_histogram(stat="count", mapping = aes(color=status)) + ggtitle("Health Status by Species (No Humans and No Aliens)") + scale_fill_hue(l=80, c=12) + scale_color_hue(l=40, c=35)

ggplotly(noHumansOrAliens)
```

# Health Status Looking at Character Type

```{r}
countHealth = newdata %>% 
  ggplot(aes(x = rickOrMorty, fill=status)) + 
  geom_histogram(stat="count", mapping = aes(color=status)) + ggtitle("Health Status by Character Type") + scale_fill_hue(l=80, c=12) + scale_color_hue(l=40, c=35)

ggplotly(countHealth)

rickAndMorty= filter(newdata, rickOrMorty != 'Neither') %>% 
  ggplot(aes(x = rickOrMorty, fill=status)) + 
  geom_histogram(stat="count", mapping = aes(color=status)) + ggtitle("Health Status for Rick and Morty Characters") + scale_fill_hue(l=80, c=12) + scale_color_hue(l=40, c=35)

ggplotly(rickAndMorty)
```

# Species and Status of Rick and Morty Characters only (bad way)
```{r}
rickAndMorty= filter(newdata, rickOrMorty != 'Neither') %>% 
  ggplot(aes(x = species, fill=rickOrMorty)) + 
  geom_histogram(stat="count", mapping = aes(color=status),position="dodge") + ggtitle("Health Status for Rick and Morty Characters by Species")+ scale_fill_hue(l=80, c=12) + scale_color_manual(values=c("#33cc33", "#000000","#9933ff"))
ggplotly(rickAndMorty)
```


# Species and Status of Rick and Morty Characters only (Better way)
Two separate graphs, instead of one overloaded graph
```{r}

rick= filter(newdata, rickOrMorty == 'Rick') %>% 
  ggplot(aes(x = species, fill=status)) + 
  geom_histogram(stat="count", mapping = aes(color=status)) + ggtitle("Health Status for Rick Characters") + scale_fill_hue(l=80, c=12) + scale_color_hue(l=40, c=35)

ggplotly(rick)

morty = filter(newdata, rickOrMorty == 'Morty') %>% 
  ggplot(aes(x = species, fill=status)) + 
  geom_histogram(stat="count", mapping = aes(color=status)) + ggtitle("Health Status for Morty Characters") + scale_fill_hue(l=80, c=12) + scale_color_hue(l=40, c=35)

ggplotly(morty)
```