---
title: "Stat. 651 ggplot2"
author: "Prof. Eric A. Suess"
output:
  word_document: default
  html_notebook: default
  pdf_document: default
---

ggplot2 examples

```{r message=FALSE}
library(tidyverse)
library(mdsr)
```

# CIACounties

Make the base plot **g** and then add different layers on to it.

```{r}
head(CIACountries)

# base plot g

g <- CIACountries %>% ggplot(aes(y = gdp, x = educ)) 

g + geom_point()

g + geom_point(size = 3)

g + geom_point(aes(color = net_users), size = 3)

# no geom_point used for the next picture

g + geom_text( aes(label = country, color = net_users), size = 3 )

g + geom_point( aes(color = net_users, size = roadways) )

```

Change the scales

```{r}
g + geom_point(aes(color = net_users, size = roadways)) +
  coord_trans( y = "log10")

g + geom_point(aes(color = net_users, size = roadways)) +
  scale_y_continuous(name = "Gross Domestic Product", trans = "log10")

```


# Faceting

```{r}
g + geom_point(alpha = 0.9, aes(size = roadways)) +
  coord_trans(y = "log10") +
  facet_wrap( ~ net_users, nrow = 1) +
  theme(legend.position = "top")

g + geom_point(alpha = 0.9, aes(size = roadways)) +
  coord_trans(y = "log10") +
  scale_y_continuous(name = "Gross Domestic Product", trans = "log10") +
  facet_wrap( ~ net_users, nrow = 1) +
  theme(legend.position = "top")
```


# Export the data and try in Tableau

```{r}
getwd()

write_csv(CIACountries, "CIACountries.csv")
```


# MedicareCharges

Check out the [MEPS](https://meps.ahrq.gov/mepsweb/) website for more real data.

```{r}
# head(MedicareCharges) # This now causes an error, remove the grouping.

? MedicareCharges

MedicareCharges <- ungroup(MedicareCharges)

head(MedicareCharges)

NJCharges <- MedicareCharges %>% filter(stateProvider == "NJ")
NJCharges
```

```{r}
p <- NJCharges %>% ggplot(aes(y = mean_charge, x = reorder(drg, mean_charge))) +
  geom_bar(fill = "grey", stat = "identity")
p

p <- p + ylab("Statewide Average Charges ($)") + 
         xlab("Medical Procedure (DRG)")
p

p <- p + theme(axis.text.x = element_text(angle = 90, hjust = 1))
p
```

Now add the overall data to the plot to compare with NJ.

```{r}
p <- p + geom_point(data = MedicareCharges, size = 1, alpha = 0.3)
p
```

# SAT

Here is the link to the [College Board SAT](https://research.collegeboard.org/programs/sat/data) website.

```{r}
g <- SAT_2010 %>% ggplot(aes(x = math))

g + geom_histogram()

g + geom_histogram(binwidth = 10)

g + geom_density()
```

```{r}
ggplot( data = head(SAT_2010, 10), aes( y = math, x = reorder(state, math) ) )  +
  geom_bar(stat = "identity")
```

Scatterplot with tend lines

```{r}
g <- SAT_2010 %>% ggplot(aes(x = expenditure, y = math)) +
  geom_point()
g

g <- g + geom_smooth(method="lm", se = 0) +
  xlab("Average expenditure per student ($100)") +
  ylab("Average score on math SAT")
g
```

Add the trend line within groups representing rate of taking the test.

```{r}
SAT_2010 <- SAT_2010 %>%
  mutate(SAT_rate = cut(sat_pct, breaks = c(0, 30, 60, 100), labels = c("low", "medium", "high") ))

g <- g %+% SAT_2010
g

g + aes(color = SAT_rate)

g +facet_wrap( ~ SAT_rate)
```

# HELPPrct

Here is the link to the [NSDUH](https://www.samhsa.gov/data/data-we-collect/nsduh-national-survey-drug-use-and-health) website.

```{r}
HELPrct %>% ggplot(aes(x = homeless)) +
  geom_bar(aes(fill = substance), position = "fill") +
  coord_flip()
```

# NHANES

Here is the link to the [NHANES](https://www.cdc.gov/nchs/nhanes/index.htm) website.

```{r}
library(NHANES)

head(NHANES)
```

Take a sample first and then make the plot.  

```{r}
sample_n(NHANES, size = 1000) %>% ggplot(aes(x = Age, y = Height, color = Gender)) +
  geom_point() +
  geom_smooth() +
  xlab("Age (years)") +
  ylab("Height (cm)")
```

Here is an alternative plot using all the data.  This is hexbin plot.

```{r}
NHANES %>% ggplot(aes(x = Age, y = Height, color = Gender)) +
  geom_hex() +
  geom_smooth() +
  xlab("Age (years)") +
  ylab("Height (cm)")
```

```{r}
library(mosaic)

head(NHANES)

NHANES2 <- NHANES %>% select(AgeDecade, BMI_WHO) 
head(NHANES2)

NHANES2_table <- table(NHANES2)
NHANES2_table

mosaicplot(NHANES2_table, color = TRUE)

```


# Weather 

```{r}
library(macleish)

head(whately_2015)

whately_2015 %>% ggplot(aes(x = when, y=temperature)) +
  geom_line(color = "darkgrey") +
  geom_smooth() +
  xlab(NULL) +
  ylab("Tempurature (degrees Fahrenheit)")
```

Here is the link to the [choroplethr](https://arilamstein.com/open-source/) website.

```{r}
library(choroplethr)
library(choroplethrMaps)
library(rUnemploymentData)

animated_state_unemployment_choropleth()

# animated_county_unemployment_choropleth()

```

# Networks

Check out [ggnet2](https://briatte.github.io/ggnet/) or the newer [ggnetwork](https://briatte.github.io/ggnetwork/)

Example 4.

```{r}
library(GGally)
library(network)
library(sna)
library(ggplot2)

# root URL
r = "https://raw.githubusercontent.com/briatte/ggnet/master/"

# read nodes
v = read.csv(paste0(r, "inst/extdata/nodes.tsv"), sep = "\t")
names(v)

# read edges
e = read.csv(paste0(r, "inst/extdata/network.tsv"), sep = "\t")
names(e)

# network object
net = network(e, directed = TRUE)

# party affiliation
x = data.frame(Twitter = network.vertex.names(net))
x = merge(x, v, by = "Twitter", sort = FALSE)$Groupe
net %v% "party" = as.character(x)

# color palette
y = RColorBrewer::brewer.pal(9, "Set1")[ c(3, 1, 9, 6, 8, 5, 2) ]
names(y) = levels(x)

# network plot
ggnet2(net, color = "party", palette = y, alpha = 0.75, size = 4, edge.alpha = 0.5)

```

Review Table 3.3 on page 47 for the different kinds of plots that can be made for different kinds of x, y variables.


Continue with the Extended example: Historical baby names on page 48.