---
title: "EDA"
output:
  flexdashboard::flex_dashboard:
    orientation: rows
    vertical_layout: scroll
    theme: flatly
    social: menu
    source_code: embed
    navbar:
      - { title: "Dataset Description", href: "#dataset-description" }
      - { title: "Univariate Analysis", href: "#univariate-analysis" }
      - { title: "Bivariate Analysis", href: "#bivariate-analysis" }
      - { title: "Multivariate Analysis", href: "#multivariate-analysis" }
---


## Dataset Description {.tabset .active} {#dataset-description}

### View of Dataset

```{r}
library(ggplot2)
library(dplyr)
library(reshape2)
library(viridis)

data <- read.csv("RHS_2011_Table_4_0.csv")

colnames(data) <- c("State", "Crude_Birth_Rate_Total", "Crude_Birth_Rate_Rural", 
                    "Crude_Birth_Rate_Urban", "Crude_Death_Rate_Total", 
                    "Crude_Death_Rate_Rural", "Crude_Death_Rate_Urban")

head(data)

```

### About the Dataset


```{r}

str(data)

```
### Summary of Dataset


```{r}
summary(data)


```

## Univariate Analysis {.tabset} {#univariate-analysis}


### Histogram for Crude Birth Rate - Total

```{r}
ggplot(data, aes(x=Crude_Birth_Rate_Total)) +
 geom_histogram(aes(y=..density..),breaks= seq(10,30,by=2),
 fill="blue",color="black",alpha=0.7) +
 geom_density(color="red",size=1) +
 ggtitle("HistogramwithDensityofCrudeBirthRate-Total") +
 theme_minimal()

```

### Histogram for Crude Death Rate - Total

```{r}
 ggplot(data, aes(x=Crude_Death_Rate_Total)) +
geom_histogram(aes(y=..density..),breaks= seq(5,10,by=0.5),
 fill="green",color="black",alpha=0.7) +
 geom_density(color="red",size=1) +
 ggtitle("HistogramwithDensityofCrudeDeathRate-Total") +
 theme_minimal()
```

### Histogram with breaks and density for Crude Birth Rate - Rural


```{r}
 ggplot(data, aes(x=Crude_Birth_Rate_Rural)) +
 geom_histogram(aes(y=..density..),breaks= seq(10,35,by=2),
 fill="purple",color="black",alpha=0.7) +
 geom_density(color="red",size=1) +
 ggtitle("HistogramwithDensityofCrudeBirthRate-Rural") +
 theme_minimal()

```


### Histogram with breaks and density for Crude Death Rate-Rural

```{r}
 ggplot(data, aes(x=Crude_Death_Rate_Rural)) +
 geom_histogram(aes(y=..density..),breaks= seq(5,10,by=0.5),
 fill="orange",color="black",alpha=0.7) +
 geom_density(color="red",size=1) +
 ggtitle("HistogramwithDensityofCrudeDeathRate-Rural") +
 theme_minimal()



```
### Histogram with breaks and density for Crude Death Rate-Urban

```{r}
ggplot(data, aes(x=Crude_Death_Rate_Urban)) +
 geom_histogram(aes(y=..density..),breaks= seq(2,8,by=0.5),
 fill="brown",color="black",alpha=0.7) +
 geom_density(color="red",size=1) +
 ggtitle("HistogramwithDensityofCrudeDeathRate-Urban") +
 theme_minimal()
```

### Histogram with breaks and density for CrudeBirthRate-Urban


```{r}
ggplot(data, aes(x=Crude_Birth_Rate_Urban)) +
 geom_histogram(aes(y=..density..),breaks= seq(10,25,by=2),
 fill="cyan",color="black",alpha=0.7) +
 geom_density(color="red",size=1) +
 ggtitle("HistogramwithDensityofCrudeBirthRate-Urban") +
 theme_minimal()
```


## Bivariate Analysis {.tabset} {#bivariate-analysis}


### Boxplot for Crude Birth Rate - Total

```{r}
ggplot(data, aes(x = "", y = Crude_Birth_Rate_Total)) + 
  geom_boxplot(fill = "lightblue") + 
  ylab("Crude Birth Rate - Total") + 
  ggtitle("Boxplot of Crude Birth Rate - Total") + 
  theme_minimal()

```

### Boxplot for Crude Birth Rate - Rural


```{r}
ggplot(data, aes(x = "", y = Crude_Birth_Rate_Rural)) + 
  geom_boxplot(fill = "lightgreen") + 
  ylab("Crude Birth Rate - Rural") + 
  ggtitle("Boxplot of Crude Birth Rate - Rural") + 
  theme_minimal()
```

### Boxplot for Crude Birth Rate - Urban

```{r}

ggplot(data, aes(x = "", y = Crude_Birth_Rate_Urban)) + 
  geom_boxplot(fill = "black") + 
  ylab("Crude Birth Rate - Urban") + 
  ggtitle("Boxplot of Crude Birth Rate - Urban") + 
  theme_minimal()

```

### Boxplot for Crude Death Rate - Total


```{r}
ggplot(data, aes(x = "", y = Crude_Death_Rate_Total)) + 
  geom_boxplot(fill = "lightcoral") + 
  ylab("Crude Death Rate - Total") + 
  ggtitle("Boxplot of Crude Death Rate - Total") + 
  theme_minimal()

```

### Boxplot for Crude Death Rate - Rural

```{r}
# Boxplot for Crude Death Rate - Rural
ggplot(data, aes(x = "", y = Crude_Death_Rate_Rural)) + 
  geom_boxplot(fill = "lightyellow") + 
  ylab("Crude Death Rate - Rural") + 
  ggtitle("Boxplot of Crude Death Rate - Rural") + 
  theme_minimal()
```

### Boxplot for Crude Death Rate - Urban

```{r}
ggplot(data, aes(x = "", y = Crude_Death_Rate_Urban)) + 
  geom_boxplot(fill = "lightpink") + 
  ylab("Crude Death Rate - Urban") + 
  ggtitle("Boxplot of Crude Death Rate - Urban") + 
  theme_minimal()
```



## Multivariate Analysis {.tabset} {#multivariate-analysis}

### Scatter plot: Crude Birth Rate - Total vs Crude Death Rate - Total


```{r}
ggplot(data, aes(x = Crude_Birth_Rate_Total, y = Crude_Death_Rate_Total)) + 
  geom_point(color = "blue", size = 3, alpha = 0.7) + 
  geom_smooth(method = "lm", se = FALSE, color = "red") + 
  ggtitle("Crude Birth Rate vs Crude Death Rate (Total)") + 
  xlab("Crude Birth Rate - Total") + 
  ylab("Crude Death Rate - Total") + 
  theme_minimal()

```

### Scatter plot: Crude Birth Rate - Rural vs Crude Death Rate - Rural


```{r}
ggplot(data, aes(x = Crude_Birth_Rate_Rural, y = Crude_Death_Rate_Rural)) + 
  geom_point(color = "green", size = 3, alpha = 0.7) + 
  geom_smooth(method = "lm", se = FALSE, color = "red") + 
  ggtitle("Crude Birth Rate vs Crude Death Rate (Rural)") + 
  xlab("Crude Birth Rate - Rural") + 
  ylab("Crude Death Rate - Rural") + 
  theme_minimal()
```

### Scatter plot: Crude Birth Rate - Urban vs Crude Death Rate - Urban

```{r}
ggplot(data, aes(x = Crude_Birth_Rate_Urban, y = Crude_Death_Rate_Urban)) + 
  geom_point(color = "purple", size = 3, alpha = 0.7) + 
  geom_smooth(method = "lm", se = FALSE, color = "red") + 
  ggtitle("Crude Birth Rate vs Crude Death Rate (Urban)") + 
  xlab("Crude Birth Rate - Urban") + 
  ylab("Crude Death Rate - Urban") + 
  theme_minimal()

```