---
title: "Climate Change Analysis"
author: "Alexander Kurochkin"
date: "04/11/2023"
output: 
  flexdashboard::flex_dashboard:
    storyboard: true
    source: embed
---

```{r setup, include=FALSE}
library(flexdashboard)
library(rnoaa)
library(dplyr)
library(ggplot2)
library(usmap)
library(tidyverse)
library(cowplot)

bdds = read.csv('C:/Users/kuroc/Documents/From Lenovo Desktop/Univer/HU/ANLY 512 - Data Visualization/Lab 2 Data/AllStates.csv')
```

### U.S. Annual Average Temperatures

```{r}
avgT=read.csv("https://www.ncdc.noaa.gov/cag/national/time-series/110-tavg-1-6-1980-2022.csv", skip=3) 
avgT$Date=as.numeric(substr(avgT$Date, 0, 4))

ggplot(avgT, aes(Date,Value)) +
  geom_line(color="black") +
  geom_point(color="black", size=0.75) +
  geom_smooth(method='lm', se=FALSE, color='red') +
  labs(title="U.S. Annual Average Temperatures",subtitle = "1980 - 2022", caption = "Source: NOAA National Centers for Environmental Information (NCEI)", x="Year", y="Temperature (F)")
```

***

For my purposes, I decided to use the U.S. climate data for the period from 1980 to 2022. 

Here we see that the U.S. Annual Average Temperatures show an upwards trend during this period.

### U.S. Billion-Dollar Disaster Events Frequency

```{r}
bdd = read.csv("https://www.nodc.noaa.gov/archive/arc0153/0209268/13.13/data/0-data/events-US-1980-2022.csv",skip=1)
bdd = bdd %>% mutate(Year = as.integer(substring(Begin.Date,1,4)))
bdd$Disaster = factor(bdd$Disaster)

ggplot(bdd, aes(x=Year, fill = Disaster)) +
  geom_bar(stat="count")+labs(title="U.S. Billion-Dollar Cost Disaster",subtitle = "1980 - 2022",
        x ="Year", y = "Frequency")
```

***

$1B+ USD disaster occurrences are on the rise as well. In 2022 there were 6 times more such events compared to 1980.

### U.S. Billion-Dollar Disasters Map

```{r, width=800}
disasters_count = bdds[order(-as.numeric(bdds$No)), ] %>% distinct(state, .keep_all = TRUE)

# Get centroids
centroid_labels <- usmapdata::centroid_labels("states")
# Join data to centroids
names(centroid_labels)[names(centroid_labels) == 'abbr'] <- 'state'
data_labels <- merge(centroid_labels, disasters_count, by = "state")

plot_usmap(data = disasters_count, values = "No", color = "black") + 
  scale_fill_continuous(low = "white", high = "red", name = "1980-2022 $1B+ Disaster Events Count", label = scales::comma) + 
  geom_text(data = data_labels, ggplot2::aes(
    x = x, y = y,
    label = scales::number(No, accuracy = 1)
  ), size = 2, color = "black") + 
  theme(legend.position = "bottom")
```

***

The most $1B+ disasters (154) affected Texas for the period between 1980 and 2022. 

Hawaii had only one event of such scale for the same period.

### U.S. Billion-Dollar Disasters Map by Deaths

```{r, width=800}
deaths_by_disaster = aggregate(Deaths ~ state, data=bdds, sum, na.rm=TRUE)

data_labels2 <- merge(centroid_labels, deaths_by_disaster, by = "state")

plot_usmap(data = deaths_by_disaster, values = "Deaths", color = "black") + 
  scale_fill_continuous(low = "white", high = "red", name = "1980-2022 $1B+ Disasters Deaths", label = scales::comma) + 
  geom_text(data = data_labels2, ggplot2::aes(
    x = x, y = y,
    label = scales::number(Deaths, accuracy = 1)
  ), size = 2, color = "black") +
  theme(legend.position = "bottom")
```

***

If to look from the perspective of fatalities, Texas fared better than 6 other states (Mississippi, Alabama, Louisiana, Georgia, Tennessee, and Kentucky).

However, I have to note that the numbers on this map are not the exact deaths by state. These are calculated as follows: total deaths per disaster that affected one or multiple states to some extent are summed up together by an affected state. For example, if 2 states were affected by the same disaster that took a total 10 lives, then both states get that 10 added to their totals.

### U.S. Billion-Dollar Disasters Map by Event Costs

```{r, width=800}
costs_by_disaster = aggregate(Total.CPI.Adjusted.Cost..Millions.of.Dollars. ~ state, data=bdds, sum, na.rm=TRUE)

data_labels3 <- merge(centroid_labels, costs_by_disaster, by = "state")

plot_usmap(data = costs_by_disaster, values = "Total.CPI.Adjusted.Cost..Millions.of.Dollars.", color = "black") + 
  scale_fill_continuous(low = "white", high = "red", name = "1980-2022 $1B+ Disasters Costs", labels = scales::unit_format(unit = "B", scale = 1e-3, sep="")) +
  geom_text(data = data_labels3, ggplot2::aes(
    x = x, y = y,
    label = scales::number(Total.CPI.Adjusted.Cost..Millions.of.Dollars., accuracy = 1, scale = 1e-3)
  ), size = 2, color = "black") +
  theme(legend.position = "bottom",legend.text = element_text(angle = 45, vjust = 0.5, hjust = 0.5))
```

***

I have to note that the numbers on this map are not the exact costs by state. The same principle applies to the calculation of costs as the one on the previous slide. For example, if 2 states were affected by the same disastrous event that had an impact of \$2B total, then both states will have $2B added to their totals. The numbers are also rounded.

We can see that Mississippi, Alabama, Louisiana, North Carolina, and Georgia were hit by the disasters that had the most financial impact.

### The Most and The Least Disastrous States

```{r, fig.width = 10}
a <- disasters_count[1:10,] %>%
ggplot(aes(x = reorder(state, No), y = No)) + 
  geom_bar(stat="identity",width = 0.75, aes(fill=state)) +
  coord_flip()+
  labs(title = "10 States with the Most $1B+ Disasters Count",
       subtitle = "1980 - 2022",
       caption = "Source: NOAA National Centers for Environmental Information (NCEI)",
       x = "State",
       y = "$1B+ Disasters Count")

b <- map_df(disasters_count, rev)[1:10,] %>%
ggplot(aes(x = reorder(state, -No), y = No)) + 
  geom_bar(stat="identity",width = 0.75, aes(fill=state)) +
  coord_flip()+
  labs(title = "10 States with the Fewest $1B+ Disasters Count",
       subtitle = "1980 - 2022",
       x = "State",
       y = "$1B+ Disasters Count")

plot_grid(a,b, align = "v", ncol = 2, rel_widths = c(1, 1))
```

***

As we noted in the previous slides, Texas has the most $1B+ disasters, while Hawaii is the best to avoid such events.
 

### Summary

The data presented shows that the U.S. has experienced an upward trend in annual average temperatures over the period under review, and this has been accompanied by an increase in the number of \$1B+ USD disaster occurrences. In 2022, there were six times more such events compared to 1980, with Texas being the most affected state, having experienced 154 of these disasters. However, when looking at the number of fatalities resulting from these events, Texas fared better than the four states to the east (Mississippi, Alabama, Louisiana, Georgia) and Tennessee, and Kentucky. Mississippi, Alabama, Louisiana, North Carolina, and Georgia were hit by disasters with the most significant financial impact. In contrast, Hawaii had only one such event during the same period and is the best state to avoid \$1B+ disasters.

Based on the presented information, I got the following questions in mind:

1. How has the observed upward trend in annual average temperatures impacted the frequency and severity of $1B+ USD disasters in the U.S. over time?
2. Are there any patterns in the geographic distribution of $1B+ disasters in the U.S. over the period under review, and if so, what might explain these patterns?
3. To what extent is the observed increase in the frequency of $1B+ disasters in the U.S. a result of climate change, and what other factors might be contributing to this trend?
4. What are some potential long-term effects of the observed increase in $1B+ disasters on the affected states and the U.S. economy as a whole, and how might these effects be mitigated?

***

References:

NOAA National Centers for Environmental information, Climate at a Glance: National Time Series, published April 2023, retrieved on April 11, 2023 from https://www.ncei.noaa.gov/access/monitoring/climate-at-a-glance/national/time-series

NOAA National Centers for Environmental Information (NCEI) U.S. Billion-Dollar Weather and Climate Disasters (2023). https://www.ncei.noaa.gov/access/billions/, DOI: 10.25921/stkw-7w73